
Micron Technology delivered an extraordinary strong quarter, reporting Q2 FY2026 revenue of USD 23.86 billion, nearly tripling y/y and beating Wall Street's USD 20.1 billion consensus by 19%. Non-GAAP EPS of USD12.20 exceeded the USD 9.31 estimate by 31%, while gross margins hit a staggering 74.9%.
Yet in overnight trading, Micron fell over -6% despite these results and strong forward guidance, as concerns around elevated CAPEX plans, geopolitical tensions (US-Iran) and the Fed’s latest commentary weighed on sentiment.
Given the inherently cyclical nature of memory, we suggest investors treat Micron as a tactical allocation, while acknowledging an upside potential of 49% with a target price of USD 680.l
Figure 1: Micron 5-year share price

Micron’s Q2 26 result
Table 1: Micron’s Q2 26 result summary
|
Q2 26 result |
(USD) |
q/q |
y/y |
|
Revenue |
23.86 billion |
75% |
196% |
|
Gross Margin |
74.90% |
+18.1 bps |
+37.0 bps |
|
Operating Income |
16.46 billion |
156% |
720% |
|
Operating Margin |
69.00% |
+22.0 bps |
+44.1 bps |
|
Net Income |
14.02 billion |
156% |
686% |
|
Diluted EPS |
12.2 |
155% |
682% |
Source: Bloomberg Finance, Micron, iFAST Compilation, Data as of 19 March 2026.
Revenue for the quarter reached USD 23.86 billion, representing a 75% sequential increase from the USD 13.64 billion reported in the first fiscal quarter and nearly tripling the USD 8.05 billion recorded in the same period a year ago.
This USD 10.2 billion q/q revenue increase is the largest in the company's history, highlighting the strength of AI-driven memory demand. The revenue growth was consistent across all business units, each reaching all-time highs.
Profitability metrics exceeded the high end of previous guidance across the board. Non-GAAP gross margin reached a company record of 74.9%, driven primarily driven by HBM pricing premiums and favourable product mix shifts. Non-GAAP diluted earnings per share (EPS) stood at $12.20, an increase of 155% q/q and 682% y/y, well above consensus.
Table 2: Business unit performance revenue
|
Q2 26 (USD bil) |
Q1 26 (USD bil) |
q/q |
Q2 25 (USD bil) |
y/y |
|
|
Cloud Memory (CMBU) |
7,749 |
5,284 |
47% |
2,947 |
163% |
|
Core Data Center (CDBU) |
5,687 |
2,379 |
139% |
1,830 |
211% |
|
Mobile and Client (MCBU) |
7,711 |
4,255 |
81% |
2,236 |
245% |
|
Automotive and Embedded (AEBU) |
2,708 |
1,720 |
57% |
1,034 |
162% |
Source: Bloomberg Finance, Micron, iFAST Compilation, Data as of 19 March 2026.
The record-breaking results were driven by strength across all business units. The Cloud Memory Business Unit (CMBU) generated 7.7 billion in revenue, up 47 q/q, reflecting continued hyperscaler investment in AI infrastructure.
The Core Data Centre Business Unit (CDBU) experienced the strongest sequential growth at 139%, supported by demand for enterprise SSDs and standard server DRAM as data centre operators modernised their fleets to support agentic AI workloads.
The Mobile and Client Business Unit (MCBU) reported USD 7.7 billion in revenue, up 81% sequentially, driven by rising memory content per device rather than unit growth. High-end smartphones are now shipping with 12GB to 16GB of DRAM to support on-device AI, while AI-enabled PCs require a minimum of 32GB. Finally, the Automotive and Embedded Business Unit (AEBU) reached a record USD 2.7 billion, driven by the accelerating deployment of Level 2+and Level 4 autonomous driving systems, which require significant increases in memory and storage capacity, up to 300GB per vehicle for L4 autonomy.
Micron’s Q3 26 guidance
Table 3: Micron’s Q3 26 guidance
|
Q3 26 guidance |
|
|
Revenue |
USD 33.5 billion ± USD 750 million |
|
Gross margin |
~81% |
|
Operating expenses |
~ USD 1.4 billion |
|
Diluted EPS |
USD 19.15 ± USD 0.40 |
Source: Micron, iFAST Compilation, Data as of 19 March 2026.
Micron's Q3 26 guidance is notably strong, with revenue guided to USD 32.8-34.3 billion, significantly above prior consensus.
Non-GAAP EPS guidance of USD18.75-19.55 nearly doubles Q2's result. Non-GAAP gross margin is expected to reach approximately 81%, a level rarely seen in the memory industry.
To contextualize: Q3 26 guided revenue of USD 33.5 billion exceeds Micron's entire fiscal year 2024 revenue of USD 25.1 billion. If Q4 maintains anywhere near this trajectory, FY26 annual revenue could approach USD 90-100 billion, up from USD 37.4 billion in FY2025 and USD 15.5 billion in the downcycle trough of FY23. Annual non-GAAP EPS could reach USD 55-65, implying a forward P/E ratio in the single digits at current prices.
Management attributed this to continued DRAM and NAND supply tightness, AI-driven demand, and favourable product mix. They noted the company can fulfil only 50-67% of demand from key customers, with HBM supply largely committed through 2026.
The entire calendar 2026 HBM supply is sold out under long-term agreements, including the first-ever five-year strategic customer agreement in the memory industry.
Table 4: HBM product pipeline as of March 2026
|
Product detail |
Bandwith |
Capacity |
Status |
|
|
HBM3E |
12-high |
1.2 TB/s |
36gb |
High volume production |
|
HBM4 |
12-high |
>2.8 TB/s |
36gb |
Volume sipments commenced |
|
HBM4 |
16-high |
>3.3 TB/s |
48gb |
Customer sampling |
|
HBM4E |
TBC |
TBC |
TBC |
Volume ramp scheduled 2027 |
Source: Micron, Bloomberg Finance, iFAST Compilation, Data as of 19 Mar 2026.
CAPEX guidance was a key concern. FY26 CAPEX was raised to above USD 25 billion (from ~USD 20 billion previously), with further increases in FY27.
This reflects multi-year capacity expansion across the US, Singapore and Taiwan, supporting long-term demand but raising concerns over future supply normalisation.
Figure 3: Micron 5y and forecasted CAPEX

Stock tumbled post-earnings beat
Despite the strong results, Micron's stock declined post-earnings, falling to roughly $432 from a $461.73 previous close. The selloff reflects a combination of elevated CAPEX expectations, macro uncertainty, and profit-taking following a strong year-to-date rally of 62% YTD gain, rather than a deterioration of fundamentals.
Figure 4: Micron share price YTD

We believe the selloff is primarily sentiment-driven
Supply remained constrained, particularly as HBM production continues to displace conventional DRAM and NAND capacity, supported elevated pricing. Despite the increasing CAPEX indicating a potentially higher supply, typically a new fab will take 2-4 years to complete, indicating the supply is unlikely to cope with the skyrocketing demand.
Additionally, we expect the ASP will continue to be stretcher given management’s latest step to change the initial LTA (long term agreement) to STA (strategic term agreement), which we think owing to the demand shortage-driven ASP hike gave the memory players more room to negotiate for a term that favour them.
On top of that, the new entrants’ threats (Chinese players for example) is negligible, where the same logic as above, new fab is tough to build within near term, while the technology of Chinese players is not able to produce HBM yet. Having said, with higher ramp up production in older generation DRAM or NAND from other players to resolve market demand shortage, it unleash the potential of the big-3 memory players (Micron, Samsung, SK Hynix) to allocate more resource into higher-margin product such as HBM.
Memory supercycle has no precedent in scale or duration
Both DRAM and NAND contract prices have surged at an extraordinary pace from the past few months, one of the largest increases on record on most of the segments including server DRAM, PC DRAM, NAND flash or NAND SSD. Below figures show the proxy of a spike in memory spot price lastly.
Figure 5 : NAND flash spot price

Figure 6: DRAM spot price

The structural driver remains AI. Data centre and AI applications are expected to account for over 50% of industry demand in 2026. HBM production now absorbs 18-28% of total DRAM wafer capacity, creating severe constraints in conventional DRAM supply.
At the same time, supply expansion remains constrained, with meaningful new capacity unlikely before 2027-2028, reinforcing the tight supply-demand dynamic.
We expect pricing momentum to moderate from 1H27 as incremental supply comes online, which may act as a headwind to share price momentum in this case.
Tariff risk is not a concern
Tariff exposure is manageable. While certain products face tariffs, strong pricing power and supply tightness have enabled cost pass-through, limiting demand impact.
The more relevant risk lies in relative competitiveness versus Korean peers ,given differing tariff exposure.
PE suggests value, yet PB signals caution given cyclical peak contidions
Micron has historically traded at 1.0–2.5x book value through full cycles. Even at the peak of the 2017-2018 memory supercycle, P/B reached only ~2.5-3.0x. The current 6.0x is the highest P/B in Micron's history by a wide margin.
P/E alone justifies a buy. P/B alone suggests caution. The truth is between them, the P/E is more reflective of forward value creation, but the P/B reminds us that memory is still cyclical, and buying at record margins carries timing risk.
We believe that the memory players are still, a cyclical business and investors should just treat it as a tactical trade instead of core holdings as it is extremely tough to gauge when the earnings peak is for current cycle.
Nevertheless, given the current scenario, we believe that we are just at the somewhere middle between the uptrend of memory supercycle.
Table 5: Forward P/B comparison of Micron and peers
|
Average |
5.83 |
|
Micron Technology Inc |
6.04 |
|
Nanya Technology Corp |
4.19 |
|
SK hynix Inc |
4.94 |
|
Samsung Electronics Co Ltd |
2.91 |
|
Winbond Electronics Corp |
4.92 |
|
Western Digital Corp |
11.99 |
Source: Bloomberg Finance, iFAST Compilation, Data as of 19 Mar 2026.
Applying PE valuation rather than PB in our analysis
From a valuation standpoint, the stock currently presents a compelling opportunity despite its significant year-to-date rally. Historically, memory stocks have traded at low single-digit multiples of peak earnings. However, the current supercycle is characterized by structural de-risking and locked-in demand through 2026, suggesting that a traditional cyclical multiple may no longer be appropriate.
For this analysis, we adopt a forward price-to-earnings framework anchored to the Gordon-Keynes (GK) justified P/E model, which derives a fair multiple from expected earnings growth, payout ratios, and cost of equity, rather than a price-to-book approach.
While P/B remains a useful cross-check for asset-heavy industrials, we believe it is structurally ill-suited as the primary valuation tool for Micron at this stage of the cycle. Book value, recorded at historical depreciated cost, materially understates the replacement value of Micron's fab network, particularly as greenfield construction costs have escalated sharply and inherently lags the rapid equity accretion occurring at current profitability levels. More critically, P/B multiples in the memory sector have historically served as better indicators of trough valuation floors than of fair value during periods of sustained earnings expansion, making them less informative when the supply-demand framework points to a multi-year upcycle.
Tactical trade, BUY with a target price of USD 680 with 49% upside potential
Table 6: Micron’s valuation
|
FY25 |
FY26e |
FY27e |
|
|
EPS |
8.29 |
56.7 |
67.8 |
|
Growth |
584.0% |
19.6% |
|
|
Current price (19 Mar 2026) |
432 |
||
|
PE |
55.6 |
8.1 |
6.7 |
|
Fair PE |
10 |
||
|
Target price |
680 |
Source: Bloomberg Finance, iFAST Compilation, Data as of 16 Apr 2026.
Our base case assumes tight DRAM and NAND supply-demand conditions persisting through at least calendar 2028, underpinned by structural cleanroom constraints across all three major producers, rising HBM trade ratios that cannibalise conventional DRAM output, and escalating AI-driven demand vectors, including inference scaling, KV cache offload, and the nascent robotics segment that management has flagged as extending visibility well beyond the current fiscal year.
Combining the above, we derive a FY27 EPS of USD 67.8 and apply a more conservative fair PE of 10x, we have computed target price of USD 680, which translates into a 49% upside potential from today’s price of $457.
As such, we reiterate a ‘Buy’ for Micron.
Key summary
Micron’s post-earnings selloff reflects the central tension in memory investing: exceptionally strong current earnings versus uncertainty around the sustainability of the cycle.
We think the pullback as sentiment-driven, but given the cyclical nature, we suggest
investors treat Micron as a tactical allocation despite the upside potential.
