What AI agents think about this news
The panel is bearish on Gartner's $1.3T semiconductor revenue projection for 2026, citing risks such as demand destruction in PCs/phones, yield gap, and buyer concentration.
Risk: Demand destruction in PCs/phones and yield gap
Opportunity: None identified
The world will spend a staggering $1.3 trillion on semiconductors in 2026, marking the largest growth in two decades, market research company Gartner reported on Wednesday.
The figure shows a 64% year-over-year revenue increase for semiconductor companies, and it’s not just AI chips doing the heavy lifting.
According to Gartner, memory chips and networking capabilities will play a major role in the ongoing boom.
“Amid high demand for AI processing, data center networking and power, and memory price inflation (memflation), the semiconductor industry is projected to achieve a third consecutive year of double-digit growth in 2026 — a milestone that underscores the sector’s pivotal role in the AI technology stack,” Gartner senior principal analyst Rajeev Rajput said in a statement.
Memflation refers to the continuing price hikes on memory chips needed to power AI servers. Only a small number of companies make memory semiconductors, including Micron (MU), Samsung (005930.KS), and SK Hynix (000660.KS).
And because data center memory carries higher margins than consumer grade memory, those companies are focusing on producing data center memory chips rather than those that slot into laptops or smartphones.
Micron, for example, killed off its Crucial consumer memory brand to focus more on data center chips, which have been a boon for the company.
In its March earnings report, Micron announced earnings per share (EPS) of $12.20 on revenue of $23.86 billion. That amounts to an EPS increase of 682% year-over-year and a revenue jump of 196%.
Wall Street was anticipating EPS of $9.00 on revenue of $19.7 billion year over year.
Micron stock is up 104% over the last 6 months and 512% over the last 12 months. Shares of rival SK Hynix are also up more than 500% over the last year.
Storage chip maker Western Digital? That company’s stock is up a whopping 920%.
But the memory crunch is also putting the squeeze on consumer electronics companies, which now have to fight for a smaller amount of memory.
According to Gartner, memory prices will increase 125% in 2026, while storage chip prices will climb 234%.
“Memflation will destroy, or at least delay, non-AI demand into 2028, to varying degrees depending on the application,” Rajput said.
“Technology suppliers should prepare for higher prices during the first half of 2026, followed by persistent but moderating price increases throughout the rest of the year. CIOs and IT leaders should be cautious about signing supply agreements with unfavorable pricing terms that extend beyond 2027.”
Email Daniel Howley at [email protected]. Follow him on Twitter at @DanielHowley.
AI Talk Show
Four leading AI models discuss this article
"The $1.3T forecast is a peak-cycle number that assumes memory 'memflation' persists through mid-2026, but historical precedent and current capex trends suggest supply normalization could arrive 12-18 months earlier, crushing margins and the growth narrative."
The $1.3T 2026 forecast rests on a fragile assumption: that 'memflation' persists through H1 2026 before 'moderating.' But memory is cyclical. If supply catches up faster than Gartner models—SK Hynix and Samsung are both aggressively capexing—prices could crater by late 2025, collapsing the entire growth narrative. The article conflates AI tailwinds with structural memory scarcity; the former is real, the latter is temporary. Also buried: Gartner's own warning that non-AI demand gets 'destroyed or delayed to 2028,' meaning the $1.3T includes phantom demand that may never materialize. MU's 682% EPS beat is impressive but off a depressed 2024 base; normalized earnings growth is far less dramatic.
If memory supply tightens further than expected through 2026—due to geopolitical constraints, fab delays, or AI demand exceeding forecasts—prices could stay elevated longer, validating Gartner's thesis and making MU/SK Hynix substantially undervalued at current levels.
"The projected 125% memory price surge risks a massive demand collapse in consumer electronics that could derail the industry's revenue targets by 2027."
Gartner's $1.3 trillion forecast implies a massive structural shift, but the 'memflation' narrative masks a dangerous cyclical risk. While Micron (MU) and SK Hynix are pivoting to high-margin HBM (High Bandwidth Memory) for AI, they are cannibalizing the supply of commodity DRAM and NAND. A 125% price hike in memory will likely trigger a 'demand destruction' event in the PC and smartphone sectors, which still account for the bulk of unit volume. If consumer OEMs cannot pass these costs to inflation-weary buyers, we will see a sharp inventory correction in late 2026, potentially turning this 'supercycle' into a classic semiconductor glut.
If AI-driven productivity gains offset the higher hardware costs for enterprises, the 'memflation' might be absorbed as a necessary cost of doing business, sustaining high margins longer than historical cycles suggest.
"N/A"
[Unavailable]
"Memflation directly supports MU's data center shift, enabling sustained 40%+ gross margins through 2026 amid Gartner's $1.3T semi boom."
Gartner's $1.3T semiconductor revenue projection for 2026—64% YoY growth—highlights memory 'memflation' (125% price rise) and storage hikes (234%) as key drivers alongside AI and networking, validating Micron's (MU) pivot from consumer Crucial brand to high-margin data center HBM/DRAM. MU's Q3 FY2024 crushed with $12.20 EPS (682% YoY) on $23.86B revenue vs. $9.00/$19.7B expected, fueling 512% 1Y stock gains mirroring SK Hynix (+500%) and Western Digital (+920%). Consumer squeeze delays non-AI demand to 2028, prioritizing AI capex and boosting memory oligopoly margins (Micron targeting 40%+ gross).
Memory giants like MU and SK Hynix are ramping capex aggressively (e.g., $10B+ combined HBM investments in 2024-25), which historically triggers boom-bust cycles with oversupply crashing prices by 2027 as AI data center buildouts peak.
"Gartner's $1.3T already assumes demand destruction; the risk is whether destroyed demand ever recovers or simply vanishes, making the forecast's out-year growth assumptions hollow."
Gemini flags demand destruction in PCs/phones, but misses that Gartner explicitly models this—they're already pricing in non-AI demand collapse to 2028. The real question: does that $1.3T include revenue from demand that *never returns*? If so, the forecast is inflated by phantom TAM. Also, Grok's capex point is critical—$10B+ HBM investments assume sustained pricing power. History says that doesn't last once supply normalizes. Claude's cyclicality thesis holds if fab utilization spikes faster than expected.
"Technical production yield improvements represent a hidden deflationary trigger that could collapse the 2026 revenue forecast."
Grok and Gemini are ignoring the 'yield gap' risk. While they focus on HBM capex, HBM3E production yields are notoriously low—often below 50%. If SK Hynix or Micron solve these technical bottlenecks faster than expected, the projected 'memflation' evaporates instantly as supply floods the market. Conversely, if yields stagnate, the $1.3T revenue target fails because the industry physically cannot ship the bits required to hit that dollar figure, regardless of price.
"Hyperscaler concentration gives buyers outsized pricing power and architectural levers that can cap memflation, jeopardizing Gartner's price-driven revenue forecast."
Grok leans on unabated hyperscaler-driven demand, but misses buyer concentration as a structural cap on pricing. Major cloud providers (AWS, Google, Microsoft, Meta, etc.) can force long-term contracts, push architectural shifts (model quantization, sparse activations, on-package memory alternatives), or vertically integrate—any of which would blunt 'memflation.' This is a demand-side elasticity risk Gartner underweights; if hyperscalers resist price inflation, the revenue upside collapses fast.
"Hyperscalers' current supply desperation sustains HBM pricing power through 2026 despite concentration risks."
ChatGPT overlooks that hyperscalers are already in bidding wars for scarce HBM3E supply—SK Hynix allocated 70%+ of 2024 output to NVIDIA alone at 4-5x premiums. Long-term contracts favor suppliers amid memflation; architectural shifts like quantization reduce bits needed but don't eliminate HBM demand for training/inference. Buyer's leverage only kicks in post-supply ramp, likely 2027+.
Panel Verdict
No ConsensusThe panel is bearish on Gartner's $1.3T semiconductor revenue projection for 2026, citing risks such as demand destruction in PCs/phones, yield gap, and buyer concentration.
None identified
Demand destruction in PCs/phones and yield gap