You’re not wrong…you’re late.

Intel just closed the research chapter on neuromorphic computing. The question is whether it can open a commercial one before the Pentagon spends $54.6 billion on drones that need exactly what Loihi does best.

On May 13, Intel (INTC) archived Lava, the open-source software framework that gave researchers access to its Loihi chips. Months later, Mike Davies, director of Intel’s Neuromorphic Computing Lab, has described Loihi 3 as being in development and positioned it for the coming era of physical AI. Intel has not published a public spec sheet for a next-generation Loihi processor or a new SDK.

That sequencing matters for traders watching INTC ahead of its Q3 earnings call on October 29. The stock closed at $102.81 on Monday, up roughly 175% year to date. Intel reported Q2 revenue of $16.1 billion and non-GAAP EPS of $0.42. Q3 revenue guidance is $15.8 billion to $16.8 billion, with a midpoint of $16.3 billion. Neuromorphic is a small but strategically loud piece of the story, but the financials are still being driven elsewhere.

Intel’s silicon roadmap depends on the same advanced lithography commitments now reshaping the broader chip industry. TSMC and Samsung have both made dated High-NA EUV production commitments that set the competitive clock for every chipmaker on an advanced-node roadmap.

Why Drones Are the Signal Worth Tracking

The Pentagon’s FY2027 budget request put $54.6 billion into the Defense Autonomous Warfare Group (DAWG), up from $225.9 million in fiscal 2026. Much of that money is framed around autonomy and operations in contested environments, including GPS-denied conditions, which is the kind of setting where low-latency, low-power processing matters.

The results are validated. Intel and ETH Zurich demonstrated adaptive horizon tracking on a drone platform with closed-loop speeds up to 20 kilohertz and 200 microseconds of visual processing on Loihi. Separately, a Science Robotics paper demonstrated a fully neuromorphic vision-to-control pipeline for autonomous drone flight running at 200 hertz on Intel’s Loihi processor, with about 0.94 watt of idle power and only a small additional draw while running the network. For a battery-powered platform navigating a jammed corridor without satellite links, that kind of power budget can translate into materially longer endurance than conventional high-power approaches.

Neuromorphic chips process sensor data asynchronously, firing only when inputs change. GPS-denied navigation leans on optical flow and visual odometry, workloads that match the sparse, event-driven signals these chips handle natively.

The Gap Between Lab and Procurement

None of this is contracted revenue. Loihi 2 is described by Intel as a research chip, and Loihi 3 remains in development without a public spec sheet. Lava’s archival removed the most visible open-access path for outside teams, even as Intel continues to describe Lava as its neuromorphic software framework. Until Intel ships and supports a clear product-grade SDK and commercial distribution model, integrators beyond Intel’s research community still face friction that can slow real procurement timelines.

That is the tension the trade sits on. Intel carries one of the most validated neuromorphic stacks among U.S. chipmakers, and the defense autonomy budget request is large. The commercialization timeline is Intel’s to control, and right now it is still hard to underwrite.

What to Watch October 29

• Gross margin guidance: Q2 came in at 41.8%. Any slip would matter for sentiment.
• Physical AI language: Intel’s August $15 billion common stock offering announcement cited physical AI and purpose-built silicon as growth vectors. If management maps neuromorphic explicitly to defense autonomy on the call, that signals a posture change.
• SDK timeline: Any specific date on a next-generation Loihi toolchain would shorten the gap between the Pentagon’s procurement appetite and Intel’s ability to address it.

The hardware case for neuromorphic drone autopilots is no longer speculative. The commercial execution case still is. October 29 is one of the few near-term moments where management can close that gap, or widen it.