BlackBerry (NYSE: BB) has nearly tripled since January, surging from roughly $3.90 to near $9.20 - up more than 150% year-to-date. The stock ran through a four-year high late last week, then jumped another 7% in extended trading on "physical AI" momentum and retail enthusiasm. The narrative is seductive. The valuation is not.
I'm taking a Hold position. The QNX turnaround is legitimate - not vaporware, not a pipe dream. But the market has re-rated this stock from a beaten-up turnaround play to a premium AI-adjacent software name, and the multiple now demands flawless execution for years to come. That is too much proof to ask before the next quarterly print.
What has actually changed
BlackBerry's business has two remaining engines: QNX (embedded software for vehicles, robotics, and mission-critical systems) and CSecure (cybersecurity and identity services). The old phone business is gone.
The QNX story is the real one. In the fourth quarter of fiscal 2026 (reported April 9), QNX revenue grew 20% year-over-year to $78.7 million - the fastest pace in quarters, up from 10% growth in the prior year. Gross margin hit 84%, and adjusted EBITDA - earnings before interest, taxes, depreciation, and amortization, a rough proxy for cash earnings - reached $21.4 million, or 27% of revenue. That is an operating profile that any software investor would respect.
Full-year QNX revenue came to $268 million, up 14% year-over-year with 83% segment gross margin. The company's total Q4 revenue of $156 million (up 10% year-over-year) included $72.5 million from CSecure, which means QNX is now roughly half the company and pulling its weight at every metric.
Company-wide Q4 adjusted EBITDA was $36.1 million (a 23% margin), and operating cash flow reached $45.6 million. CFO Tim Foote told reporters the company has "transformed from a cash-burning business into a profitable software company." On the numbers, he's right.
The AI narrative - real, but not yet priced into revenue
Here is where the story gets more nuanced. Foote and management are positioning QNX as a critical enabler of "physical AI" - the category of AI that runs on real-world machines like autonomous vehicles, robots, and medical devices where, as they put it, "there is zero margin for error."
QNX is embedded in more than 215 million vehicles globally. The platform partners with NVIDIA on autonomous driving. The software is real, the customer base is real, and the addressable market in robotics and autonomous systems is genuinely expanding.
But "physical AI" is still a forward-looking narrative. The 20% QNX growth in Q4 was driven by royalties and development revenue from existing automotive customers - not a sudden AI-driven explosion. Management says it plans to step up investment in QNX for physical AI, robotics, and medical applications. Plans are not revenue. Not yet.
The valuation problem
This is the part that should give buyers pause. Let me lay out the numbers.
The stock's EV/EBITDA multiple surged to approximately 57x in the latest quarter - up from roughly 17.7x just two quarters earlier and far above the 10-year median of 22.7x. The forward non-GAAP P/E sits near 55x. Revenue growth for the full company was 3% for the year and 10% in the most recent quarter.
To put that in plain language: investors are paying 55 times earnings for a company growing total revenue in the low double digits. That is a multiple befitting a 30%+ growth software name with visible path to scale. BlackBerry is not there yet.
Yes, QNX margins are exceptional at 84% gross and 27% EBITDA. But QNX represents roughly half of $549 million in annual revenue. The company is still small, and the valuation assumes that QNX accelerates meaningfully while CSecure doesn't drag the picture down. Any stalling in automotive licensing cycles or any margin pressure at CSecure creates a mismatch between what the stock costs and what the business delivers.
The meme layer
BlackBerry has been a meme stock since the 2021 Reddit-fueled short squeeze. Short interest sits at approximately 5.8% of the public float (34 million shares as of mid-May). That's not exceptionally high by meme standards, but enough that a sharp move can trigger short covering and amplify the rally - or a pullback can reverse the dynamic.
The retail-driven energy around the stock is real. Reddit communities and trading forums are again buzzing. But meme momentum is a force multiplier, not a fundamental. When the narrative cools, the operating numbers have to carry the stock. At a 55x P/E, there is not much cushion if the story loses steam.
The catalyst clock
BlackBerry reports Q1 fiscal 2027 earnings on June 23 - roughly three weeks away. That is the next hard test. The market consensus expects EPS of roughly $0.03, but what matters more is QNX revenue growth, guidance trajectory, and whether management can show AI-related wins moving from pilot deals to paid contracts.
If QNX growth stays at or above 20% and guidance is raised, the current multiple might hold. If growth decelerates or management can't point to tangible AI monetization, the stock has a long way to fall from these levels - there is no earnings floor at 55x P/E when the business is growing at 10%.
Risks
- QNX deceleration: Automotive licensing is cyclical. A softening in vehicle production or a delay in next-generation software-defined vehicle programs would hit QNX first and hardest.
- CSecure drag: The cybersecurity division has been under pressure. If it continues to underperform, it pulls down the overall growth picture and justifies a smaller-multiple valuation.
- Multiple reversion risk: At 57x EV/EBITDA, the stock is pricing in years of flawless execution. A single quarter of mixed results could compress the multiple back toward the 30-35x range, which would mean a 30-40% price decline.
- Narrative fade: "Physical AI" is a powerful story, but the market has been burned by AI narratives that don't show up on the P&L. If QNX can't demonstrate AI-driven revenue growth in the next two quarters, the premium evaporates.
Investor takeaway
I would not sell shares here if you hold them - the business turnaround is too real to panic out of, and the June 23 earnings print could justify a further pop. But I would not buy at these levels either.
The stock has already rewarded patience. Investors who bought below $4 got a three-bagger from a legitimate margin and cash flow turnaround. The question now is whether the growth trajectory can justify a premium software multiple, and the evidence isn't there yet.
Wait for the June 23 earnings print. If QNX growth accelerates to mid-20% or higher, guidance is raised, and AI-driven wins show up in actual contracted revenue - not just pilot announcements - the stock earns its rerating. If growth stays in the low teens and the "physical AI" story remains aspirational, the valuation has run well ahead of the business.
Rating: Hold. Too rich to buy, too real to sell.