Meanwhile, the US Department of Defense (DoD) appears to be doing the opposite. The Chief Digital and Artificial Intelligence Office (CDAO) recently axed its Chief Technology Officer (CTO) directorate, a move many analysts view as strategic self-sabotage.

This directorate, responsible for overseeing more than $340 million in AI and digital integrations in fiscal year 2024, represented a critical nexus linking battlefield innovations with institutional infrastructure. Its elimination, justified under “efficiency” mandates, alarmed defense observers who fear it fractures continuity, erases institutional memory, and sends a dangerous signal to adversaries willing to exploit perceived American weakness.

The Strategic Misstep

The CDAO was formed in 2022 by fusing key functions from the Joint Artificial Intelligence Center, Defense Digital Service, Chief Data Office, and Advana analytics, aiming to unify policy, technology, and digital services. Embedded within CDAO, the CTO led cross-functional teams in AI, cyber, logistics, and command-and-control systems, ensuring that new technologies remained interoperable and aligned with warfighter requirements.

Abruptly dismantling this directorate not only removes a pivotal vision and coordination role but also creates a void with no clear replacement. The result is fragmented efforts, lost synergy across mission areas, and a battlefield advantage handed to adversaries.

Expertise Lost, Momentum Undermined

Leadership and expertise take years, even decades, to develop. Figures like Bill Streilein, former CTO of CDAO and veteran of MIT Lincoln Laboratory, carried institutional memory and high standards into Pentagon AI programs. But when top-tier professionals are sidelined under the label of “streamlining,” they often leave and seldom return.

This pattern has already occurred. The Defense Digital Service (DDS), once lauded as the Pentagon’s “SWAT team of nerds,” lost almost all of its members by May 2025, prompting its demise. Nearly every DDS member, citing bureaucratic pressure from the Department of Government Efficiency (DOGE), chose to depart rather than conform.

These departures are not benign transfers. They represent the scattering of core innovators and connectors whose insight and trust networks are irreplaceable. Without them, emerging AI systems risk becoming siloed projects rather than battlefield-enabling capabilities.

DOGE: Efficiency or Engineered Evisceration?

DOGE, instituted by a presidential executive order in January 2025, is authorized to slash perceived inefficiencies across federal agencies—often through AI-enhanced, automated assessments. Under the leadership of figures tied to Elon Musk and Vivek Ramaswamy, DOGE has repurposed its mandate to aggressively target leadership and innovation roles across the board—including in national defense.

DOGE has justified cuts using its proprietary AI systems to flag and eliminate “inefficient” programs, often without human oversight or contextual nuance. The CTO’s directorate was among its most high-profile targets, methodically identified and removed, despite its mission-critical nature.

To make matters worse, DOGE is reportedly comfortable with these decisions. One Pentagon official described it as a “theater of dominance,” not just cost-cutting, but deliberate erasure of institutional anchors to obfuscate the depth and breadth of the sacrifice.

The High-Stakes Fallout

Adversaries feast on the narrative that the US champions AI yet purges its own tech leadership overnight. “America cannibalizes its talent while claiming leadership in AI warfare,” such narratives go. These optics weaken American deterrence, erode allied confidence, and provide cover for Moscow, Beijing, and Pyongyang to reframe the battlefield narrative.

Domestic consequences are equally grim. The consistent removal of flagship tech roles projects a clear message to science, technology, engineering, and math (STEM) professionals; serve, and risk being discarded. That weakness is a recruitment boon for adversaries, national lab contractors, and tech-armed autocracies solving tomorrow’s warfare puzzles.

Real efficiencies lie not in gutting leadership but in fortifying it. Per the National Security Commission on AI, prioritizing disciplined recruitment and retention of technical talent, including a Digital Corps and AI fellowships, is key to American competitiveness. Instead, we witness the dismantling of precisely those anchor roles meant to shepherd AI innovation into combat-relevant systems.

The DOGE-Driven Dismantling of Tech Leadership

The concepts herein are alarming and reflect an institutional unraveling that directly undermines America’s global security posture and strategic deterrence in five critical ways. First, the elimination of the Chief Technology Officer (CTO) directorate from the Chief Digital and Artificial Intelligence Office (CDAO) strips away a core pillar of the Pentagon’s ability to adapt emerging technologies for battlefield advantage. This directorate was not redundant bureaucracy; it was the crucible in which ideas from national labs, industry, and warfighters were harmonized into operational capability.

By abruptly dismantling this team, the Department of Defense has extinguished a pipeline of institutional memory and strategic insight at the precise moment when rapid, informed, and integrated decision-making is needed. This brain drain parallels a historical pattern of self-sabotage and leaves adversaries uncontested in the tech talent race.

Second, the removal of high-level AI leadership is a propaganda gift to revisionist powers like China and Russia. These states are watching America voluntarily decapitate its own strategic leadership, an act they can now frame as proof of American decline. This strengthens their strategic messaging in influence campaigns aimed at allies, neutral states, and even American citizens.

“America cannibalizes its talent while claiming leadership in AI warfare” is not just a phrase, it is a weaponized narrative that demoralizes partners and emboldens adversaries to challenge American dominance in contested domains like cyberspace, space, and AI warfare.

Third, strategic deterrence hinges on credible capability and the perception of cohesion. DOGE’s algorithmic-driven targeting of leadership roles without contextual assessment introduces chaos into the acquisition and integration life cycle of military AI systems. Instead of creating synergistic effects across logistics, cyber, and command and control, the US risks building a fractured, siloed ecosystem that fails in joint operations.

By removing the very leaders who prevent stove piping, the US sabotages its ability to develop and field interoperable, scalable, and warfighter-ready AI tools. This systemic breakdown makes deterrence brittle, vulnerable to being cracked in future high-end conflicts.

Fourth, the US has struggled to compete with the private sector for AI and cybersecurity talent. By signaling that even elite government technologists are disposable under the guise of “efficiency,” this policy drives future talent away from public service. Those who might have joined a modern “Digital Corps” will instead seek stability and respect elsewhere, perhaps even abroad.

Strategic deterrence depends not only on weapons but on technologists who know how to deploy them. Gutting these roles ensures that tomorrow’s innovations will not make it past the lab, let alone onto the battlefield.

Fifth, DOGE’s use of automated assessments to eliminate “inefficiencies” without human oversight is a grotesque parody of reform. Its reliance on cold, context-blind algorithms to purge critical roles mimics adversary models of techno-authoritarianism, not democratic accountability. If allowed to continue, this will hollow out innovation across government agencies and military branches.

Efficiency is not the enemy, misapplied efficiency is. Strategic deterrence requires smart investments, not cost-cutting theater that sacrifices our warfighting edge on the altar of political optics.

Strategic Self-Sabotage Must Be Reversed

This is not merely streamlining, it is full-blown surrender. The dismantling of the CDAO’s CTO directorate and the broader DOGE initiative represents an engineered unraveling of the very leadership needed to project U.S. strategic deterrence in the AI era. Leadership is the vector through which technology becomes capability. Remove it, and you hand your adversaries not only the advantage, but the narrative.

Unless reversed, these concepts and actions will echo through wargames, deterrence failures, and battlefield losses. The US must stop cannibalizing its competitive edge and re-center its national security strategy on strengthening, not sidelining, its AI leadership.

Conclusion

Leadership is not just overhead on the funding spreadsheet; these leaders are our ammunition in the fight for global AI dominance. Removing them during a strategic inflection point is not reform, it is a self-made vulnerability, and as the US disables its own leadership of advanced technologies, it is dismantling future readiness.

The nation must insist on accountability. Cost-cutting means nothing if it costs the technological coherence to compete in tomorrow’s battles. In the strategic competition unfolding now, leadership is the weapon, and ceding it is surrender. This page out of the DOGE handbook should be shredded and burned. Remember, Iranian nuclear scientists were not dismantled by their own regime, they were destroyed by US and Israeli bombs.

Greg Sharpe is Marketing Director at the National Institute for Deterrence Studies. He is retired from the US Air Force. The views expressed are his own.

The Quiet Dismantling of America’s AI Warfighting Edge was originally published on Global Security Review.

In Jacobsen’s scenario, North Korea unexpectedly launches a limited preemptive strike against the American homeland. The United States responds with a decapitating strike against North Korea’s nuclear arsenal. Through unlikely errors, Russia believes that the United States is attacking them instead of North Korea. In response, Russia launches a large-scale strike. This leads the United States to launch everything. Bing, bang, boom, and we have Armageddon followed by nuclear winter and the death of billions. In interviews with Vanity Fair and Mother Jones, Jacobsen argues that the US has a plan to conduct nuclear warfighting, based on a series of previously adopted requirements that leave the president little leeway to modify his response. This problem is only made worse by the fact that the president has sole authority to employ nuclear weapons. To this point, Jacobsen argues that military “war mongers” have a very “aggressive culture” that has a predilection to “jam” the president toward “quickly launching a massive retaliatory strike.”

Jacobsen asserts that American intercontinental ballistic missiles are on “HAIR TRIGGER ALERT.” Why? She interviewed William Perry who used the term. Bam, nuclear weapons are on hair trigger alert and missileers have itchy trigger fingers.

While admitting that nuclear deterrence has worked for seven decades, Jacobsen warns that deterrence will fail and when it does, any use of a nuclear weapon will result in large-scale nuclear war. Why? Jacobson interviewed Paul Bracken, who participated in government sponsored wargames—four decades ago—that ended in general nuclear war. Ipso-facto, the use of one nuclear weapon leads to Armageddon.

In Jacobsen’s scenario, American missile defenses are a complete failure. Why? She interviewed Ted Postol who said they do not work—case closed. Her conclusion is that missile defense is a farce and a waste of money. And the United States is lying to itself if it thinks they will ever stop an inbound nuclear weapon.

Elsewhere in the scenario, the president is forced to launch a large-scale nuclear response against the Russians within six minutes. Why? Jacobsen was also told by Perry that the president “will not wait” and thus assumed that Ronald Reagan’s memoir, in which he made an off-hand remark about having six minutes to decide on nuclear weapons use, reflects policy. Abracadabra, the president must decide to use nuclear weapons within six minutes.

Jacobsen’s predilection for incorrectly contextualizing the statements of those she interviewed is both stunning and worrying. The conclusions she draws about the present from the statements of former, often Reagan era or earlier, officials boggles the mind. Jacobsen does not seem to grasp the fact that she or her interviewees may be wrong about the relevance of their past experience to the present. Three examples are indicative.

Problems

First, Jacobsen is wrong about the alert status of the nation’s intercontinental ballistic missile force. They are not on “HAIR TRIGGER ALERT.” The United States does not have a launch-on-warning or launch-under-attack policy/doctrine. Nothing of the sort exists. Adam Lowther and Derek Williams dismantle this argument and explain that the United States maintains a launch-under-attack option, which allows the president to employ intercontinental ballistic missiles pre-, mid-, or post-strike. The option requires nothing of the president.

A number of variables will influence this decision. What is important to remember is that there is tremendous work that goes into thinking through scenarios well before they ever arise. Thus, the idea that these weapons are on hair trigger alert is ridiculous. Years of planning and analysis take place left of launch.

Second, Jacobsen is clearly unfamiliar with the design and purpose of wargames. They are specifically designed to understand the implications of a concept or capability. This means a scenario is artificially designed to ensure participants achieve the game’s objectives. Thus, when Jacobsen assumes that because a wargame or series of wargames end in general nuclear war, that a real conflict must necessarily end in general nuclear war, she is fundamentally misunderstanding the purpose and arbitrary nature of wargames. Within the wargaming community, it is well understood that they are not predictive of the future but are instructive of potential options.

Third, missile defenses are not worthless, as Jacobsen claims. They are making steady improvements in their ability to destroy targets. Ted Postol, her primary source of information about missile defenses, was wrong about the effectiveness of Israel’s Iron Dome system. Hard data is proving that missile defenses, in this case Iron Dome, are far more effective than Postel believed.

Although there are additional areas where Jacobsen incorporates inaccurate information into her scenario, the point is clear. Annie Jacobsen’s Nuclear War: A Scenario would be far more accurately titled, Nuclear War: A Novel or Nuclear War: Disarmament Propaganda. The biggest challenge with the book is that Americans with little understanding of nuclear operations will believe the bias with which Jacobsen writes. This makes it imperative that those within the nuclear community speak out and correct the record. Nuclear deterrence is too important to turn over to a journalist with an agenda.

Peter Huessy is a Senior Fellow at the National Institute for Deterrence Studies. The views expressed in this article are the author’s own. 

Annie Jacobsen Gets It Wrong about Nuclear Deterrence was originally published on Global Security Review.