India's smartphone market just posted its worst June quarter in six years — and the culprit isn't a slowdown in demand but a chip-factory pivot in South Korea that nobody in Dharavi or Dehradun voted for.
Picture a Croma store on a busy Saturday afternoon in Bengaluru. A college student — first-generation smartphone upgrader, budget firmly fixed at ₹12,000 — is turning over a handset, checking the specs. The sales rep leans in: "Sir, same phone was ₹10,999 last December." The student puts it back on the shelf and walks out. Multiply that walk-out by tens of millions across India, and you have the story of Q2 2026.
India's smartphone shipments declined 10% year-on-year in Q2 2026 (April–June), marking the biggest decline for a June quarter in six years, according to Counterpoint Research's Monthly India Smartphone Tracker. That's not a rounding error. India ships somewhere north of 35–37 million smartphones in a typical June quarter. A 10% drop means roughly 3–4 million fewer devices changed hands — devices that would have been calls made, businesses transacted, children educated.
And the cause has almost nothing to do with India itself.
Every smartphone, no matter how cheap or expensive, needs two types of memory: DRAM (Dynamic Random Access Memory — the chip that keeps your open apps running smoothly, like the RAM in a laptop) and NAND flash (the storage chip that holds your photos, messages, and apps even when the phone is off). Together, their cost feeds into the BoM — the bill of materials, the ingredient list for any device.
For years, this ingredient list was relatively stable. Memory prices went through cycles — boom, bust, stabilise — roughly every three or four years, driven by how much Samsung, SK Hynix, and Micron (the three companies that collectively make nearly all the world's DRAM) chose to build. It was a rough oligopoly, but a somewhat predictable one.
Then generative AI arrived and turned the whole thing upside down.
AI model training and inference — the process of running a large language model to answer your query — is extraordinarily memory-hungry. The preferred chip for this isn't your phone's ordinary DRAM. It's a specialised, stratospherically expensive type called HBM (High Bandwidth Memory), where multiple DRAM dies are stacked vertically like floors in a skyscraper and connected through microscopic channels called through-silicon vias. This stacking technique dramatically shortens data transfer paths and enables a 1,024-bit ultra-wide interface, delivering memory bandwidth far beyond what conventional DRAM can achieve.
The voracious demand for HBM by hyperscalers — Microsoft, Google, Meta, Amazon — has forced the three biggest memory manufacturers to pivot their limited cleanroom space and capital expenditure toward higher-margin enterprise-grade components. This is a zero-sum game: every wafer allocated to an HBM stack for an Nvidia GPU is a wafer denied to the LPDDR5X module of a mid-range smartphone or the storage chip of a consumer laptop.
The numbers on who controls this market are stark. According to IDC, SK Hynix ranked second in the overall DRAM market (including HBM) with a 29.1% revenue share in Q1 2026, and first in HBM specifically with a 56.4% revenue share. Add Samsung and Micron, and you have three players accounting for well over 90% of the market. SK Hynix itself acknowledges that its strengths in HBM, server DRAM, and enterprise SSDs are what it relies on to navigate the memory cycle — meaning the pivot is deliberate, not accidental.
Three companies. One market. No meaningful competition from outside. And all three have been quietly diverting their fabs away from the chips that go into your ₹12,000 phone. Counterpoint Research's Tarun Pathak put it directly: "A lot of these memory companies are asking smartphone vendors to stand in line behind the hyperscalers."
The chain reaction from Korean and American fabs to Indian retail shelves was swift and brutal.
TrendForce's survey reveals that conventional DRAM contract prices increased approximately 93–98% quarter-over-quarter in Q1 2026 — the largest quarterly increase ever recorded. In Q2 2026, conventional DRAM contract prices were projected to rise another 58–63% QoQ, as suppliers continued reallocating capacity toward HBM and server applications. NAND flash contract prices were forecast to rise 70–75% QoQ in the same quarter — outpacing DRAM's rate of increase for the first time in this cycle.
At the component level, DDR5 chip prices rose from $6.84 in September 2025 to $27.20 by December — a near-quadrupling in just three months. DDR4 spot prices soared alongside, at points trading above DDR5 in some configurations, inverting the traditional value hierarchy and reflecting the growing scarcity of legacy memory.
In cumulative terms, smartphone memory prices have increased nearly 4x since September 2025 and are expected to rise further, potentially reaching 5x in the coming months.
For context: the raw material in your phone's memory slot went up in price the way petrol did during the 1973 oil shock — except this shock is structural, not geopolitical.
India is not a premium market. It is a volume market — a vast, price-sensitive ocean where the ₹8,000–₹15,000 segment has traditionally carried the bulk of shipments and where a ₹500 price difference is the difference between a sale and a walk-out.
Rising DRAM and NAND prices increased memory's share of the bill of materials from below 20% to over 45% in the sub-₹15,000 segment. Think about that for a moment. Nearly half the cost of building a budget phone is now just the memory inside it. Brands faced two choices: absorb the cost (and bleed) or pass it on (and lose customers). Most chose the latter.
Senior Analyst Prachir Singh of Counterpoint noted that "India's smartphone market remained under pressure during the quarter, as both demand and supply were adversely affected. On the supply side, persistent increases in memory and other component costs prompted almost every major OEM to implement multiple rounds of price hikes, resulting in an average smartphone price hike of around 15% by the end of the second quarter."
Smartphone OEMs implemented multiple price increases in 2026, with the highest increase exceeding 100% of the launch price. Not a typo — some models doubled in price from their original launch.
The result was a segment-level catastrophe. The mass-market segment (sub-₹15,000) was the hardest hit, with its shipments declining 45% year-on-year.
As most Chinese brands are heavily exposed to the entry- and mid-tier segments, their overall market share fell to its lowest level for a second calendar quarter since 2020. In response, many OEMs expanded their 4G portfolios in the mass-market segment to better address changing market conditions.
That last point is almost ironic. India spent years racing to 5G. The memory crunch is pushing brands to dust off their 4G playbooks.
Not everyone bled equally. Q2 2026's brand dynamics told a revealing story about who is insulated and who is exposed.
Vivo (excluding iQOO) led India's smartphone market in Q2 2026 with an 18% share, driven by strong momentum in the premium segment following the launch of the V70. Among the top five smartphone manufacturers, Samsung was the only one that registered growth — 2% year-on-year. Samsung's relative resilience owes something to an uncomfortable fact: it is both a memory chip maker and a smartphone brand, giving it supply-chain priority that pure-play phone brands simply don't have.
Apple's iPhone shipments declined 3% year-on-year in Q2 2026, with its market share reaching 7%. While consumer demand for the iPhone 17 series remained strong, persistent supply constraints and inventory shortages across online and offline channels limited the brand's shipment growth. Even Apple — which has long-term supply agreements and a war chest to pre-purchase components — couldn't fully escape the crunch.
The brightest spot was, counterintuitively, a small British startup. Nothing was India's fastest-growing smartphone brand in Q2 2026 at 105% year-on-year, driven by strong demand for the Phone (4a) and Phone (4a) Pro models, and enhanced brand visibility from its title sponsorship of the RCB cricket team during the IPL. This marked the ninth time in the last ten consecutive quarters that Nothing emerged as India's fastest-growing smartphone brand — a remarkable run, though it is worth noting the company is growing from a very small base, and its aggressive IPL marketing likely pulled it above a tide that was swamping everyone else.
On the premium end, Google emerged as the fastest-growing smartphone brand in the ultra-premium segment (above ₹45,000) in Q2 2026, posting 68% year-on-year growth — partly because, unlike rivals, it held its Pixel prices steady and quietly expanded offline availability.
One structural shift worth noting: India's smartphone market is increasingly running on credit. Smartphone financing through NBFCs and credit/debit card EMI accounted for over 50% of mainline smartphone sales in India during Q2 2026. The ultra-premium segment remained relatively resilient, supported by the growing adoption of financing options — these schemes reduced the upfront cost of premium smartphones, making them more accessible despite broader pricing pressures.
This is the paradox of the moment: the market that's hurting most is the cash buyer at the lower end. The person who can afford to borrow — and pay 18–24% annualised interest on an EMI — is still buying. The person who saves up cash to buy a phone outright is priced out. The memory crisis is, perversely, accelerating India's smartphone debt economy.
There is a case for cautious optimism, though it comes with important caveats.
Memory prices will keep climbing through Q3 2026, but the blistering increases of the past few quarters are set to cool as consumer buyers reach the ceiling of what they can afford. TrendForce projects conventional DRAM contract prices to rise 13–18% QoQ in Q3 2026, with NAND flash contract prices increasing 10–15% — substantial gains, but a marked slowdown from the roughly 60% jumps recorded in Q2.
The cooldown is driven by consumer electronics manufacturers' unwillingness and inability to absorb higher memory costs after months of relentless price increases, rather than by improved supply. In other words, memory remains in short supply, but consumers are no longer willing to keep paying ever-higher prices. This is demand destruction doing what price signals are supposed to do — cold comfort for a brand that has already lost 45% of its volume.
The hard constraint: meaningful new capacity expansion is unlikely until late 2027 or 2028. Building a new semiconductor fab takes three to four years and costs $15–25 billion. No one is turning that tap on overnight. Suppliers are expected to rely primarily on process migrations to expand bit output in 2026, with wafer input growth remaining limited to incremental gains from manufacturing optimisation.
Counterpoint expects India's smartphone market to decline 13% year-on-year for the full year 2026. Since component prices are unlikely to normalise before next year, affordability will remain the industry's biggest challenge.
Here is the deeper story, the one that matters beyond the quarterly data: India's smartphone market — around 150 million units shipped in 2025, the world's second-largest — does not control its own fate. It is downstream of decisions made in Icheon, South Korea (SK Hynix HQ), Suwon (Samsung), and Boise, Idaho (Micron) about where to direct their wafer capacity. Those decisions are increasingly being made in favour of Nvidia's data-centre customers and against the Infinix buyer in Patna.
The memory market in 2026 is being reshaped by AI servers and high-performance computing. Memory is no longer behaving like a simple consumer-electronics cycle; it is becoming an allocation-driven market. Independent assessments show this tightness is structural, not cyclical, with supply expected to remain constrained through 2027 as AI workloads continue to outpace wafer expansions.
India has no domestic memory chip manufacturer. There is no Samsung Electronics India producing LPDDR5X in Chennai. Every DRAM chip in every phone sold in every Reliance Digital store in the country is imported, priced in dollars, and subject to the production priorities of three foreign companies optimising for AI margins.
This is a supply-chain vulnerability that the Production Linked Incentive (PLI) scheme — which has successfully brought smartphone assembly to India — hasn't touched. Assembly is a downstream activity. The upstream oligopoly remains entirely out of Indian hands.
The AI boom has given India's smartphone market an unexpected bill. A market that was celebrating record shipments last year is staring at a 13% annual contraction in 2026 — not because Indians stopped wanting phones, but because the world's AI data centres want the same chips more, and can pay a lot more for them.
The student who walked out of that Croma store will upgrade eventually — maybe on EMI, maybe next year when prices stabilise, maybe on a refurbished device. But the broader point stands: India's digital ambitions now have a structural dependency on a three-player chip oligopoly that is rationally, legally, and quite profitably making a different set of priorities its own. That is the real number no quarterly tracker captures.