NVIDIA abruptly announced last night that Q2 results would fall well short of prior guidance — a genuine negative surprise.
NVIDIA's fiscal calendar differs from most: Q1 covers February–April. By the time it reported Q1 in late May, the market had already experienced the dual demand-supply shock in consumer electronics since the Shanghai lockdown. Numerous PC OEMs reported unprecedented order cancellations; consumer demand fell off a cliff.
Yet NVIDIA still guided for only a modest sequential decline and 24% year-over-year growth in Q2. Even that modest sequential deceleration led the market to deem Q1 a miss.
Intel then missed first, posting its first loss since 2009. AMD held up, delivering 25% year-over-year growth in Client, but its discrete GPU business was dismal and guided for further decline in Q3. At the time I did not grasp the severity.
My prior view on PCs was that despite macro weakness, high-end demand (gaming laptops, business laptops) remained robust. Meanwhile, NVIDIA GPUs enjoyed near-monopoly share because AMD's discrete GPU SKU lineup was thin. Reality proved me wrong: I severely underestimated the 'collapse' in consumer GPU purchase intent.
With PC and smartphone markets this weak, why were AMD, Qualcomm, and Apple relatively resilient in Q2? I believe the growth logic differs. A common thread across AMD, Qualcomm, and Apple this quarter: market share gains.
AMD: Clear share gains in high-end notebook CPUs; CPU share up for the ninth consecutive quarter
Qualcomm: Samsung flagship Snapdragon share rose from 40% to 75%; Honor, the only domestic brand with high shipment growth; Apple
Apple: Clear share gains in emerging markets; Brazil revenue up high double digits; Indonesia, Vietnam, India revenue nearly doubled
NVIDIA's GPU share was not eroded by AMD or Intel, but the overall discrete GPU market is depressed; NVIDIA overestimated its alpha. Discrete GPUs are high-end but less essential than CPUs. The RTX 30 series is at the end of its product cycle, and everyone knows RTX 40 will deliver a massive performance leap, so wait-and-see sentiment is intense. Compounding this, the crypto crash flooded the channel with used mining cards masquerading as new, further sapping new-card demand. This underscores that in the current macro, B2B businesses are far more stable than B2C.
NVIDIA took a $1.3B inventory write-down in Q2, representing 41% of Q1 inventory — reminiscent of the 2018 crypto bust. Last time, Gaming took three quarters to recover; this time, is management ripping the band-aid off?
Notably, recent Q2 earnings from major U.S. game publishers all missed; Sony and Microsoft gaming also disappointed. PC/console gamer enthusiasm has visibly waned. As an upstream chip supplier to gaming, NVIDIA faces significant pressure; we'll see what Q3 guidance looks like.
Long term, NVIDIA still has a very bright outlook; I remain confident. Next year, its data center CPU efforts could break new ground.