Here is what actually happened.

On May 27, HP reported Q2 fiscal 2026 earnings. Revenue came in at $14.4 billion, up 9% year over year, against a Street consensus of roughly $13.99 billion. Non-GAAP diluted EPS was $0.86, against an estimate of $0.72 and above the company’s own guidance range of $0.70 to $0.76. The stock jumped 10.1% the next session. Four days later, Nvidia named HP a launch partner for the RTX Spark chip at Computex 2026. HPQ gained another 8.45% on June 1. That is over 18% in a week for a company most people still describe as a printer and laptop vendor.

What’s interesting is that neither move came from thin air.

The Q2 beat was not lucky timing. HP had pre-positioned low-cost inventory ahead of commodity price increases, reconfigured products to use alternative components, and applied granular pricing adjustments by customer, geography, and channel. That is a level of supply chain management that most investors were not modeling for a company in HP’s category. Meanwhile, the Personal Systems segment delivered $10.2 billion in revenue, up 13% year over year, with operating profit growing 30% YoY even as total units shipped fell 7%. Revenue grew because AI-capable devices carry significantly higher average selling prices than standard configurations. The unit volume decline actually tells you something useful: the mix is shifting toward premium, and HP is capturing the margin differential.

The Print segment posted $4.2 billion in revenue, flat year over year, with an operating margin of 18.3%. Consumer printing was down 10%, commercial held flat. The part worth watching is the All-In Plan subscription business, which grew double digits. That is recurring revenue in a segment most people are still mentally writing off as a declining annuity. They may be wrong about the trajectory, or at least about the speed of it.

Full-year non-GAAP EPS guidance was revised upward to $2.90 to $3.10. At roughly $29.33 per share, that is a forward multiple of approximately 9.5x to 10x earnings. For context, that is still trading at a meaningful discount to the broader hardware sector.

Slight tangent, but it matters here: HP’s AI PC mix jumped from 35% to 44% of total shipments in a single quarter. Management’s own guidance projects that reaching 60% to 70% by fiscal 2027 and above 70% by fiscal 2028. That is not speculation. It is baked into their forward guidance before a single RTX Spark device has shipped.

Now the Nvidia piece, because this is where it gets interesting.

At Computex 2026, Jensen Huang introduced RTX Spark: an Arm-based superchip combining a Blackwell GPU with a 20-core Arm CPU, up to 128GB of unified memory, and 1 petaflop of AI performance. HP was named a launch partner alongside Dell, Lenovo, Microsoft, ASUS, and MSI. HP’s OmniBook was specifically highlighted as one of the thinnest RTX Spark laptops in the initial lineup, with availability expected in fall 2026. The positioning matters: Nvidia is building RTX Spark as a local agentic AI platform, meaning it runs AI agents without cloud dependency. That is a fundamentally different value proposition than anything in HP’s existing product catalog. And when chip architecture shifts this significantly, the installed base of enterprise devices does not age gracefully. It ages out. That obsolescence cycle is the demand driver HP is sitting in front of over the next 18 to 24 months.

The risk side of this is real and worth being direct about.

The strategic low-cost inventory that made Q2 margins look strong will diminish through Q3 and Q4. Memory and storage input costs are expected to keep rising through the back half of fiscal 2026. Management’s own Q3 non-GAAP EPS guidance is $0.61 to $0.71, a meaningful step down from Q2’s $0.86, and the CFO explicitly flagged Q4 as a potential margin low point before recovery into fiscal 2027. HP is also operating under interim CEO Bruce Broussard following the departure of Enrique Lores in February 2026, with a permanent search still ongoing. Leadership transitions during critical product cycles carry execution risk that does not show up in EPS models. The RTX Spark revenue itself will not hit HP’s financials until fiscal Q1 2027 at the earliest. And the PC unit TAM is projected to decline at a high-teens rate in the second half of calendar 2026, per the company’s own commentary. None of this kills the thesis. But holding it loosely matters.

The options market has already placed its read.

Going into May 27 earnings, HPQ’s weekly implied volatility reached 156 against a 52-week IV range of 26 to 61. The call-to-put ratio was running approximately 2.7 to 1, with concentrated activity in the $28 weekly calls. That positioning proved well-calibrated. Then on June 1, Barchart’s Unusual Options Activity flagged nearly 90 times the normal call volume at the July 17 expiry $35 strike. At a premium of roughly 92 cents, the breakeven sits at $35.92, which is a 23.2% move from the June 1 close. That is not a hedge. At that distance out-of-the-money, with that kind of volume concentration, it is a directional bet on momentum carrying through mid-July. Whether it is right is a separate question. But someone with size thought it was worth making.

The FCF picture gives that positioning some rational grounding. Trailing twelve-month levered free cash flow stands near $3.28 billion. At the current market cap of approximately $26.8 billion, that is an FCF yield around 14%. That is unusually high for a technology company, at any point in the cycle.

Here is where I am at on the structure of the trade.

For those looking at HPQ directionally: a call debit spread in the July or August expiry, centered around the $30 to $35 range, captures upside from continued AI PC mix expansion while capping capital at risk to the net debit. On the other side, if you think Q3 and Q4 margin compression hits harder than guided and the stock fades into the earnings setup, a put debit spread in the $25 to $27 range for August or September reflects that view with defined exposure. Traders who think the stock stays rangebound through summer, say $27 to $33, and want to sell into elevated IV could look at an iron condor structure at those wings. Post-event IV compression would be the driver there, not direction. These are analytical frameworks. Position sizing relative to total portfolio exposure matters more than any individual entry level.

HP is up 32.6% year-to-date and trading near a 52-week high of $29.55. At roughly 10.8x trailing earnings, it is not priced for collapse. It is also not priced for the kind of margin recovery the RTX Spark cycle would need to deliver to sustain a move above $35. The gap between the current price and the strike being bought in size is the unresolved question. Not an answer.

Fall 2026 is when we find out.