Palantir, Oracle, and Microsoft: Defense AI Becomes a Structural Spending Cycle

Palantir, Oracle, and Microsoft are emerging as the clearest public-market beneficiaries of the Pentagon’s accelerating appetite for artificial intelligence, and that is exactly why Dan Ives’ latest call is drawing so much attention. The Wedbush analyst’s framing is simple but powerful: defense AI is no longer a niche procurement theme; it is becoming a structural spending cycle that could reshape software and cloud winners for years. In that stack, Palantir is the first mover, Oracle is the infrastructure layer, and Microsoft is the enterprise and platform heavyweight. The implication is bigger than one quarterly beat or one budget memo — it is that AI is moving from pilot projects to mission-critical defense architecture.

Background​

The investment case for defense AI did not appear overnight. It grew out of several overlapping trends: the military’s modernization drive, the intensifying competition with China, the commercialization of generative AI, and a broader federal push to streamline AI procurement. In April 2025, the White House said the Office of Management and Budget was revising federal AI-use and procurement policies to remove bureaucratic restrictions and speed adoption across agencies. That policy shift matters because it signals that the government is trying to normalize AI as an operational tool rather than a science project. (whitehouse.gov)
By June 2025, the Department of Defense was already putting hard numbers behind the trend. It released the fiscal 2026 Military Intelligence Program top-line request at $33.6 billion, aligning the request with the Interim National Defense Strategic Guidance. Even though the detailed program breakdown remained classified, the public topline is enough to show that intelligence modernization is still receiving serious capital allocation. That is the budget environment in which software vendors are now competing for share. (defense.gov)
The broader market also offers context. Oracle has turned AI demand into a major cloud backlog story, with remaining performance obligations climbing to $523 billion in fiscal Q2 2026 and cloud infrastructure revenue growing 68% year over year. Microsoft, meanwhile, reported Azure and other cloud services growth of 34% in fiscal 2025, with the company repeatedly emphasizing AI capacity expansion and commercial bookings momentum. Palantir has paired government roots with unusually fast growth, reporting Q4 2025 revenue of $1.407 billion, up 70% year over year, alongside a sharp rise in U.S. government revenue.
That is why Ives’ comments are resonating now. His thesis is not merely that these companies sell “AI.” It is that defense AI will require a stack, and that the stack will be divided among companies with different strengths: operational software, cloud infrastructure, and enterprise integration. That is a much more investable idea than the vague promise that every defense contractor will somehow benefit from machine learning.

What Ives Is Really Arguing​

At the center of Ives’ view is the idea that defense AI spending will become a recurring budget line, not an isolated experiment. That distinction matters because markets often reward pilot-stage enthusiasm, then punish it when the work does not scale. Here, the opposite may be true: the more AI gets embedded into command, logistics, analysis, and procurement, the more durable the revenue opportunity becomes. That is the reason bulls keep calling this a generational cycle.

A stack, not a single winner​

Ives’ naming of Palantir, Oracle, and Microsoft reflects a layered view of the market. Palantir is positioned closer to the mission and workflow layer, Oracle is closer to the underlying cloud plumbing, and Microsoft brings the enterprise identity, productivity, and cloud ecosystem that can stitch systems together at scale. That means the defense AI opportunity is not zero-sum in the way many stock narratives are. A single program can create demand across multiple vendors.
The stack framework also helps explain why investors should resist the temptation to treat “defense AI” as a one-stock story. In practice, the government does not buy a monolith; it buys platforms, integrations, hosting, model access, and secure applications. That gives large incumbents and specialized software firms room to coexist, provided they can prove security, reliability, and integration value.

Why the timing is different now​

The recent inflection is tied to two forces: policy and capability. On the policy side, federal agencies are being pushed to adopt AI more aggressively and buy it more efficiently. On the capability side, the models themselves have crossed a threshold where they can actually support workflows at scale. That is an important difference from prior waves of “digital transformation,” which often stalled in pilot purgatory.
The geopolitical backdrop also adds urgency. Modern warfare increasingly depends on faster sensor fusion, better targeting, quicker logistics decisions, and more adaptive intelligence analysis. In that environment, the question is not whether AI will be used — it is which vendors will be trusted to run it in systems where latency, security, and accountability matter more than consumer convenience.

• Defense AI is becoming procurement reality, not just strategy language.
• Policy changes are lowering friction for agency adoption.
• Mission-critical workflows reward vendors with proven government credibility.
• The opportunity spans software, infrastructure, and enterprise integration.

Palantir: The First Mover​

Palantir is the most obvious beneficiary because it has spent years building trust with government clients that demand secure, auditable, high-friction deployments. Its Gotham platform is already embedded in intelligence and defense workflows, which is why Ives sees it as the first mover. That first-mover advantage matters more in government than in consumer tech because switching costs are higher, procurement is slower, and mission confidence is everything.
The numbers support the thesis, at least at a high level. Palantir said U.S. government revenue grew to $1.85 billion in 2025, up 55% year over year, and total Q4 2025 revenue reached $1.407 billion, up 70%. The company also guided FY2026 revenue to roughly $7.18 billion to $7.20 billion, implying about 61% growth. Those are extraordinary numbers for a company already as large as Palantir has become.

Why government matters more than ever​

Palantir’s government mix is not a weakness in this environment; it is a moat. Defense and intelligence customers value continuity, security accreditation, and operational fit more than they value the cheapest software license. That means once Palantir is inside a workflow, the company can often expand within that account rather than fight for replacement business. This is the kind of retention story investors pay a premium for.
The company is also increasingly able to frame AI as an operational multiplier rather than a feature add-on. That is important because agencies are not just buying model access; they are buying decision support, data fusion, and workflow execution. Palantir’s pitch is that it can turn AI into action, which aligns closely with defense use cases.

Valuation is still the tension​

The bull case is clear, but the stock still asks a demanding question: how much growth is enough? Palantir trades at a lofty multiple relative to more mature software peers, and that valuation means even excellent execution can leave impatient investors wanting more. The market is effectively pricing Palantir as if it will keep compounding at a near-hypergrowth rate for longer than most software companies ever manage.
That creates a strange paradox. The stronger the operational narrative becomes, the more the stock can still wobble if growth merely meets expectations rather than smashes them. In other words, Palantir may be winning the business case while still fighting the multiple case.

• Palantir’s moat is government trust and workflow depth.
• Its revenue mix gives it direct exposure to defense modernization.
• The company is already proving scale, not just aspiration.
• Its valuation remains the main source of investor friction.

Oracle: The Infrastructure Layer​

Oracle’s role in the defense AI story is less flashy but arguably just as important. If Palantir is the mission software, Oracle is the cloud foundation underneath it. Oracle Cloud Infrastructure has been growing quickly on the back of AI demand, with fiscal Q2 2026 cloud infrastructure revenue at $4.1 billion, up 68% year over year, and total RPO at $523 billion. Those numbers suggest customers are signing up for long-duration capacity commitments well before the revenue appears in the income statement.
That backlog matters because AI infrastructure is capital intensive and capacity constrained. Defense customers care not only about performance, but also about sovereignty, reliability, and secure workloads. Oracle has been positioning itself as a destination for serious AI infrastructure, including multicloud and database-centric deployments that can support regulated environments. That gives it a credible route into government-adjacent workloads without needing to be the face of every end-user application.

Why Oracle’s backlog is strategically important​

RPO is the cleanest symbol of Oracle’s AI momentum because it captures demand that has not yet shown up in current revenue. A $523 billion backlog is a statement that enterprises and AI companies are reserving capacity far into the future. In defense, where planning cycles are long and deployment risk is high, that kind of pipeline can be especially valuable.
Oracle’s opportunity is also asymmetric. Unlike pure software vendors that must sell directly into every workflow, Oracle can benefit when others build on top of its infrastructure. If defense AI platforms need a secure cloud layer, Oracle can capture that spend whether the visible application is Palantir, a contractor solution, or a bespoke government stack.

The market is still underestimating the plumbing story​

Investors often prefer the application layer because it is easier to understand. But in AI, the plumbing is where a lot of economic value can accumulate, especially when workloads are large, persistent, and mission-critical. Oracle’s forward multiple remains far more modest than Palantir’s, which is one reason some investors view it as the cleaner risk-adjusted way to play the theme.
At the same time, Oracle is no longer just a legacy database story. Its AI and cloud positioning has become central to the equity narrative, and that shift is still being absorbed by the market. If the defense AI spend cycle accelerates, Oracle may emerge as one of the most underappreciated beneficiaries.

• Oracle captures the infrastructure spend beneath defense AI.
• Large RPO is a strong signal of future monetization.
• Its cloud role gives it leverage without requiring application dominance.
• The stock may still look cheaper than the growth story deserves.

Microsoft: The Enterprise Backbone​

Microsoft is the most broad-based beneficiary because it brings together cloud, identity, productivity, developer tooling, and AI integration. Its Azure and other cloud services grew 34% in fiscal 2025, and the company continues to describe AI capacity expansion as a core investment priority. Microsoft also has long-standing federal relationships, which matters when the buyer is the Pentagon or a related agency that values familiar procurement pathways.
This is the subtle power of Microsoft’s role in defense AI. It does not need to own the entire mission stack to be indispensable. It just needs to sit at enough points of control — cloud, identity, collaboration, model access, application hosting — that it becomes hard to displace. In a defense environment, that kind of systemic embed is often more valuable than a single killer app.

Azure and the defense adjacency​

Azure’s growth is the headline, but the deeper story is the ecosystem. Microsoft can connect front-office and back-office systems with AI workflows, while also providing enterprise-grade security and compliance. That combination is particularly useful in government, where AI adoption often depends on whether the vendor can meet security requirements without forcing a separate technology program.
Microsoft’s relationship with OpenAI also gives it a model-layer advantage, though the exact economics are more complex than a simple ownership narrative. The important point is that Microsoft can offer a broad AI platform, not just one service. That makes it a natural partner for agencies that want to experiment with AI now but avoid being locked into a narrow toolset later.

Scale, not novelty, is Microsoft’s edge​

Microsoft’s challenge is not proving relevance; it is proving that AI can continue expanding inside a business already this large. That is usually harder than it sounds. Yet the company has repeatedly shown that it can layer new growth on top of its existing base without wrecking the core franchise.
In defense AI, that matters because procurement buyers often prefer stable vendors with broad product ecosystems. Microsoft’s biggest advantage may be that it can be introduced into an account for one reason and then expand across many others. In that sense, it is the least speculative of the three names Ives highlighted.

• Microsoft combines cloud, productivity, and security depth.
• Its enterprise footprint makes defense adoption easier.
• AI integration can spread across multiple products and workflows.
• It is the most diversified of the three plays.

The Defense AI Budget Context​

The budget backdrop is what transforms this from a trading theme into a long-duration thesis. The White House has already said federal AI procurement should be faster and less bureaucratic, and the Defense Department has publicly released the fiscal 2026 Military Intelligence Program topline at $33.6 billion. Those two facts, taken together, suggest AI modernization is not a side project — it is becoming part of the institutional machinery of government. (whitehouse.gov)
That matters for vendors because defense and intelligence budgets often create durable revenue streams once a platform is embedded. These programs can persist for years, with incremental expansions, renewals, and adjacent use cases. That is why investors are increasingly treating defense AI as a budget category rather than a single contract event.

What procurement reform changes​

Procurement reform may sound dull, but it is central to the thesis. When AI buying becomes easier, vendors with existing federal credibility gain an advantage because they can move faster than new entrants trying to prove compliance from scratch. That favors companies already known to the government ecosystem.
It also changes the competitive dynamic versus traditional defense primes. The old model assumed hardware and contracting relationships would dominate. The new model suggests software, cloud, and data platforms could claim a larger share of mission value. That is a meaningful redistribution opportunity.

Why classified budgets still matter to public investors​

Even when line items stay classified, the direction of spending still shapes the market. A rising intelligence topline implies more room for digital modernization, secure compute, analytics, and AI-assisted decision systems. Investors do not need every detail to understand the likely beneficiaries.
That said, hidden budgets can also create false confidence. Without visibility into subcategories, it is easy to overread the scale of any one vendor’s opportunity. The smart approach is to look for evidence in revenues, bookings, backlog, and contract announcements rather than assuming every defense dollar lands evenly.

Competitive Implications​

The most interesting consequence of Ives’ thesis is that it reorders the competitive landscape. Instead of asking which company “wins AI,” investors should ask which company captures which layer of the defense stack. That frames Palantir, Oracle, and Microsoft as complementary beneficiaries in some cases, but also as competitors for platform mindshare.
This will likely pressure other large vendors to sharpen their own defense AI stories. Traditional defense contractors will try to protect their budgets by bundling software more tightly with hardware and systems integration. Hyperscalers will keep arguing that cloud scale and model access are the real bottlenecks. Specialized AI vendors will claim domain expertise.

Rivals cannot ignore the software shift​

What makes the current cycle unusual is that software is not just sitting on top of defense modernization; it is increasingly the modernization. That means vendors that fail to present a credible AI story risk being treated as legacy providers, even if their hardware or services remain essential.
The competitive danger is especially high for firms that can’t articulate how their offerings improve decision speed, reduce labor, or increase battlefield awareness. In a budget environment constrained by politics and urgency, “we also do AI” is not enough. Buyers want proof of operational leverage.

Market consequences beyond defense​

There is also a broader market implication. If defense AI proves durable, it could validate a larger national-security software sector and unlock more private capital for regulated AI infrastructure. That may encourage startups, systems integrators, and cloud specialists to target government contracts more aggressively.
At the same time, the concentration of spending among a few favored vendors may lead to richer multiples for the leaders and weaker sentiment for everyone else. When the market decides a small number of firms own the core of a theme, capital tends to concentrate fast.

• The market is shifting from “who owns AI?” to “who owns each layer?”
• Legacy vendors face pressure to prove AI relevance.
• Defense primes may need software partnerships to stay competitive.
• Capital is likely to concentrate around proven winners.

Valuation, Sentiment, and Investor Behavior​

The investment case is not only about fundamentals; it is also about sentiment. Palantir can post extraordinary growth and still draw skepticism because its valuation leaves little room for disappointment. Oracle can surprise on backlog and still be treated as a transition story. Microsoft can deliver steady AI progress and still be viewed as too large to re-rate aggressively.
That creates three very different stock behaviors around one theme. Palantir is the momentum name with the biggest emotional swings. Oracle is the “show me the monetization” name. Microsoft is the institutional compounder that often attracts lower drama but also less explosive multiple expansion.

Why retail and institutional investors diverge​

Retail investors often respond to narrative acceleration, especially when a company appears to be at the center of a major platform shift. That helps explain the volatility around Palantir. Institutions, by contrast, often care more about earnings durability, margin structure, and contract quality, which favors Microsoft and sometimes Oracle.
Sentiment also matters because defense AI is the kind of theme that can become self-reinforcing. Once a consensus forms that a few names are the canonical winners, every new budget headline or contract rumor can trigger incremental buying. The danger is that the trade can get ahead of the actual revenue recognition curve.

Multiple compression is the hidden risk​

Investors in high-growth names often focus on sales growth while underestimating how quickly a multiple can compress if rates move, execution slips, or guidance disappoints. That is especially relevant for Palantir, where expectations are already very high. Oracle and Microsoft may be less vulnerable to abrupt multiple shocks because their valuations are more grounded in mature cash-flow profiles.
That does not mean they are safer in every sense. It means the market has different standards for each. The same contract announcement can be explosive for one stock and merely confirmatory for another.

Strengths and Opportunities​

The strongest version of the thesis is that the defense AI market creates multiple winners with different functions, and the companies named by Ives are already positioned to benefit from real spending, not speculative hype. The opportunity is broad enough to support a multi-year investment case, especially if federal procurement keeps moving toward faster AI adoption.

• Palantir has the deepest government workflow credibility.
• Oracle has a compelling infrastructure and backlog story.
• Microsoft brings enterprise scale, cloud breadth, and procurement familiarity.
• The defense AI stack creates room for multiple revenue streams from one spending cycle.
• Federal policy is moving toward faster AI adoption and procurement.
• The opportunity is likely to extend beyond the current budget cycle.
• The theme supports both growth investors and quality compounder investors, depending on risk tolerance.

Risks and Concerns​

The bullish narrative is strong, but the trade still carries real hazards. The biggest is that investors may extrapolate too much from policy direction and not enough from procurement reality, which can lag for years. There is also the risk that competition compresses margins or that one vendor’s momentum is more visible than its actual share of end-market spend.

• The government procurement cycle can be slow, opaque, and politically sensitive.
• Classification limits make it hard to measure true program breadth.
• Valuation risk is substantial, especially for Palantir.
• Execution risk remains high for any vendor scaling defense workloads.
• Rivals may respond with aggressive bundling and pricing.
• AI infrastructure demand could eventually face capacity or cost constraints.
• Sentiment can reverse quickly if one company misses expectations.

Looking Ahead​

The next phase of this story will be less about rhetoric and more about evidence. Investors should watch for contract awards, backlog expansion, federal deployment references, and whether the companies named in this thesis continue converting AI demand into visible revenue. The most important test will be whether defense AI shows up not just in press releases, but in recurring financial disclosures.
The other thing to watch is whether the stack remains as cleanly divided as it looks today. In a fast-moving market, platform boundaries can blur quickly, and the same vendor may move from infrastructure into application layers or vice versa. That is why the real competitive question is not who owns AI in the abstract, but who becomes indispensable at the point where software, cloud, and mission operations intersect.

• New federal AI procurement guidance and implementation detail.
• Additional Defense Department budget signals for FY2026 and beyond.
• Fresh contract awards or task orders involving the three companies.
• Whether Oracle and Microsoft keep expanding AI backlog and cloud utilization.
• Whether Palantir can sustain growth without multiple compression.

For now, Ives’ thesis captures the market’s most important defense AI insight: the opportunity is real, the buyers are institutional, and the winners may not all look the same. If the Pentagon’s AI modernization effort continues to mature, Palantir, Oracle, and Microsoft could each occupy a different rung of a very large ladder — and that may be exactly where the enduring value gets built.

---

Source: 24/7 Wall St. Wedbush's Ives: Palantir, Oracle, Microsoft to dominate defense AI integration