BrainChip Flags Akida 1500 Yield Problems and Cash Concerns Even as Defense and Medical Wearable Orders Build
Q2 2026 earnings call, September 10, 2026
BrainChip Holdings used its fiscal second-quarter update to walk investors through a strategy that CEO Sean Hehir says is finally converging with market demand, but the call also surfaced two issues that will likely dominate investor conversations in the near term: manufacturing yield problems on the company's flagship Akida 1500 chip and a cash position that is drawing sharper scrutiny as burn continues.
Yield Issues on Akida 1500 Add a New Execution Risk
The most consequential disclosure came in response to a shareholder question about chip yields. Hehir confirmed that Akida 1500 silicon is coming in below expectations, a manufacturing issue the company is still diagnosing. "There are many variables that come to a manufacturing process that can impact the yield process," he said, adding that BrainChip and its design and manufacturing partners are "analyzing every element of this to find exactly what it is." He noted the earlier multi-project wafer run had "very high yield," implying the problem emerged specifically in the production run, not the underlying design. Management declined to give a timeline beyond promising more detail at the Q3 update, leaving investors with an open-ended overhang on a product line the company has positioned as a near-term revenue driver. This is a tangible, quantifiable risk that didn't exist in prior updates and deserves closer tracking, particularly given BrainChip's reliance on 1500 shipments to seed its module and reference-design pipeline.
Cash Position Draws Shareholder Pressure
With $20 million in cash as of June 2026, funding strategy was the single most repeated shareholder question, submitted by six separate investors according to IR head Troy Dahms. Hehir's answer offered no new capital raise announcement but confirmed the board has set explicit cash-flow neutrality targets tied to the annual strategic plan. "We have line of sight and a plan at some point to make the company cash flow neutral," he said, while acknowledging that "as and if we need capital raises in between," management will bring options to the board. No specific runway figure, raise size, or timeline was disclosed, which will keep dilution risk front of mind for shareholders given the stock's history of capital raises.
No Bookings Target Met Last Year, Executive Bonuses Zeroed Out
Hehir disclosed that BrainChip missed its bookings goal in the prior year, resulting in materially reduced bonuses for both employees and executives. "We did not achieve our bookings goal last year, so the bonus for the employees and the executives were substantially down... my individual goal was only on bookings, or my incentive. We did not get it. So I got no bonus," he said. This is a rare moment of unscripted candor on a call otherwise built around confidence in the strategy, and it underscores that despite the commercial narrative, the company has not yet converted its expanded product funnel into hard bookings growth. Management also declined to provide any forward financial targets, describing the business as still "lumpy" and saying guidance will only come once there is more predictability in order flow.
Defense and Medical Wearables Remain the Clearest Commercial Traction Points
On the positive side of the ledger, Hehir pointed to defense and wearables as the two verticals generating the most concrete deal activity, driven by structural shifts in both markets. In defense, he argued that the shift from fixed to mobile platforms is forcing customers toward edge inference solutions that cannot rely on network connectivity, a dynamic playing out in BrainChip's radar and drone work with Raytheon and the Air Force Research Lab, where the company's chips add classification capability radar systems previously lacked. In wearables, ultra-low power consumption is the primary purchase driver. Hehir highlighted Onsor, a company building seizure-prediction glasses for epilepsy patients, saying the device can predict a seizure "within one hour with 98% accuracy," a use case he called emblematic of the freedom-of-life impact the technology can enable. On the hardware side, defense contractor Parsons has placed an initial order of several thousand Akida 1500 chips as part of what Hehir described as "the first of a multiyear partnership," a concrete data point on real, ongoing hardware demand even as the yield issue clouds supply.
Hybrid Compute and the IP-Plus-Silicon Business Model
Much of the call was devoted to reinforcing BrainChip's strategy of selling both licensable IP and physical chips rather than choosing one path, a model Hehir defended directly against a shareholder question about whether the company was abandoning IP licensing in favor of silicon. "IP and chips together and modules is critical, fundamental to our well-planned strategy," he said, noting that large customers often buy chips first to prototype before committing to a custom IP license, creating what he called "a very supportive virtuous circle." The company's IP catalog is expanding from a single offering to four tiers over the next several months, and its first major IP licensee, Edge AI, remains on track to tape out a chip later this year. The Akida 2500 tape-out is also progressing, according to Hehir, though no new timeline specifics were given. On the hardware side, roughly 25% of the initial Akida 1500 chip inventory has already sold, with remaining stock being partly redirected into modules meant to make adoption easier for smaller customers unable to design their own silicon.
The hybrid compute theme, in which Akida chips work alongside CPUs and GPUs rather than replacing them, was framed as a validating trend rather than a new product direction. Hehir pointed to the IBM Symphony community bundle as a proof point of this architecture gaining traction, while sidestepping a shareholder question about why BrainChip isn't doing more directly with IBM's Kevin Johnson, framing the company's role as building "the fundamental building blocks" that partners like IBM then extend into applications, rather than building end-use cases itself.
Management Pushes Back on Shareholder Frustration
Dahms directly raised shareholder discontent with company performance and leadership, to which Hehir responded by acknowledging impatience while defending the pace of execution. "I understand frustrations with shareholders because people like things quicker... but we're highly confident we're doing the right plan, and the execution is going incredibly well," he said. Combined with the disclosure of missed bonus targets and the unresolved yield issue, the tone suggests a company still asking investors for patience while working through concrete production and commercial execution problems, rather than one ready to point to hard financial inflection.