The company delivered a broad beat and raise, driven by AI-agentic server CPU demand and improved supply/yield execution, which positively inflected the narrative.
CPU numbers positioned the company directly in line with one of the market's hottest themes, and commentary regarding yields and packaging was much better.
Management indicated that Q1 revenue would have been significantly higher if supply had been sufficient, with unmet demand starting with a 'B'.
Supply is expected to increase every quarter through 2026, supported by wafer starts, yield, cycle-time improvements, and external foundries.
Management emphasized that AI workloads are shifting towards inference, agentic, and multi-agent use cases, where CPUs serve as the orchestration/control-plane layer, making the CPU structurally more important in AI infrastructure.
DCAI revenue grew 22.4% year-over-year, with management stating that server CPU demand improved over the last 90 days and is expected to drive double-digit industry and company unit growth in 2026, extending into 2027.
ASIC revenue was up over 30% quarter-over-quarter and nearly doubled year-over-year, with the business already at an annualized run rate exceeding $1 billion.
Q1 gross margin benefited from higher volume, mix, pricing, better 18A yields, and sales of previously reserved inventory.
18A yields are ahead of internal targets, and 14A maturity/yield/performance are outpacing 18A at a similar stage, with design commitments expected in 2H26 to 1H27.
Advanced packaging demand was framed as scaling into billions annually, and Terafab with SpaceX/xAI/Tesla adds high-profile optionality.
CCG beat Street expectations, and AI PCs now constitute over 60% of client CPU mix.
The company expects client revenue to hold up better than the unit TAM due to pricing and customer inventory replenishment, despite an expected low double-digit decline in the full-year PC unit TAM.
The company has re-entered the AI narrative through server CPUs, a part of the stack that investors were underestimating and now desire more exposure to.
Management argues that inference, reinforcement learning, agentic AI, physical AI, and edge AI require significantly more CPU orchestration than traditional training clusters.
The quarter provided financial support for the thesis, with DCAI up 22.4% year-over-year, guided up double digits quarter-over-quarter for Q2, multiple long-term agreements including Google, Xeon 6 selected as host CPU for NVIDIA DGX Rubin NVL8, and ASIC revenue already at a $1B+ annualized rate.
Bulls view this as evidence that the company's x86 franchise is not just stabilizing but becoming a capacity-constrained AI infrastructure asset.
The second bull pillar is execution and supply, with management stating that demand exceeds supply across businesses, 18A yields are ahead of internal targets, 14A is tracking better than 18A at a comparable point, and supply should increase every quarter.
The narrative has shifted from concerns about the company's survival and fab funding to how quickly it can increase supply to meet demand.
The capital expenditure mix shift, with tool spend up 25% year-over-year despite total capex being flat, indicates that prior shell investments can be monetized through equipment, yield, and cycle-time improvements.
Foundry optionality adds another layer, with external design commitments expected in 2H26/1H27, advanced packaging potentially scaling into billions annually, and Terafab offering a prominent U.S. manufacturing optionality.
Some bulls project a 2030 EPS power of ~$8, assuming the company captures over one-third of a much larger AI CPU market and builds a ~$25B+ foundry business, leading to a stock price over $120 at 25x P/E.
Despite the strong quarter, the stock has already priced in a significant turnaround before the most challenging aspects are proven.
The company remains a high-capex, lower-margin, execution-sensitive IDM with substantial foundry losses, a soft PC market, and rising input costs.
While Q1 gross margin was excellent, bears note that part of the upside came from previously reserved inventory, and Q2 gross margin is expected to step down despite revenue guidance being well above Street expectations.
Concerns exist that 18A remains below corporate-average margin, Foundry is still losing billions, opex is increasing, NCI is a headwind, and gross margin leverage may be lower than expected due to pricing being absorbed by memory, substrates, T-glass, wafers, and other cost inflation.
Bears question the sustainability of server CPU upside once supply normalizes and competition intensifies, with AMD remaining a strong x86 competitor and ARM-based options from NVIDIA, AWS, and Google posing a medium-term share risk.
Management is attempting to accelerate Coral Rapids and add multithreading, but bears argue that the company is still closing product gaps rather than clearly leading.
On PCs, the guidance assumes client revenue can remain roughly flat after Q2, even though FY26 PC units are expected to decline by low double digits, which relies on pricing and customer inventory replenishment in a weakening market.
Bears compare the company's expected 10%-15% sales CAGR to faster AI semi peers growing 30%-40% and argue that a 40x-60x P/E on CY26/CY27 EPS is difficult to justify until external foundry wins and durable margin recovery are visible.
PC softness and 18A/input-cost margin headwinds are noted as minor concerns.
Q2 gross margin steps down to 39% because 18A/Panther Lake ramps materially, inventory benefits do not repeat, and 18A remains below corporate-average margin.
Opex moving towards ~$16.5B, higher interest, NCI of ~$250M per quarter, and rising input costs are the main reasons some earnings estimates moved less than revenue estimates.