Bitget App
Trade smarter
Buy cryptoMarketsTradeFuturesEarnSquareMore
Delivery period exceeds 19 weeks, prices see the largest monthly increase—could “chipflation” become the next life-or-death test in AI trading?

Delivery period exceeds 19 weeks, prices see the largest monthly increase—could “chipflation” become the next life-or-death test in AI trading?

智通财经智通财经2026/07/21 07:12
Show original
By:智通财经

Behind the sharp fluctuations in stock prices, the simultaneous acceleration of delivery times and price increases seems to indicate that a deeper structural contradiction is emerging—"Chipflation."

According to English Financial News, investment institution Susquehanna stated in a recent report that the global semiconductor industry’s lead times further extended in June, a trend that remains significant despite rising prices.

Analyst Christopher Rolland noted that the semiconductor industry’s delivery time in June saw the biggest month-on-month increase of this cycle, rising by 5 days to 19.4 weeks. Even more striking, industry pricing in June registered the "largest single-month increase," up 5% from the previous month. The simultaneous acceleration of lead times and price increases highlights that the supply-demand structure for chips continues to tighten.

Additionally, the increase in June’s lead times was "broad-based"—about 81% of the companies covered had stable or rising lead times, and all distributors saw growth. Lead times for all product categories rose month-on-month; Rolland believes this "indicates the upcycle is now expanding beyond analog parts."

By category, the power discrete device sector remains tight, with lead times for power ICs and MOSFETs increasing by more than 10 days. FPGAs are also in short supply, with Lattice Semiconductor (LSCC.US) and Xilinx under AMD (AMD.US) experiencing an expansion in lead times by two weeks for the fifth month in a row. The supply of passive components is also tightening rapidly, with both Vishay (VSH.US) and Murata Manufacturing (MRAAY.US) seeing longer lead times.

At the company level, ON Semiconductor (ON.US), Diodes (DIOD.US), Renesas Electronics (RNECY.US), and ROHM Semiconductor (ROHCY.US) in Japan saw notable surges in lead times, whereas Texas Instruments (TXN.US), Microchip Technology (MCHP.US), and Infineon (IFNNY.US) remained "generally stable." Skyworks (SWKS.US), MaxLinear (MXL.US), and Coherent (COHR.US) also experienced "substantial increases" of over 10 days.

The Semiconductor Sector Experiences a “Rollercoaster” Ride, “Chipflation” May Become the Next Big Test for AI Trades

In stark contrast with the ongoing supply-demand constraints, US chip stocks saw severe turbulence in July. The Philadelphia Semiconductor Index dropped around 17% during July, although it’s still up 65% year-to-date. Last week alone, the index slid nearly 10%, the biggest single-week decline in over a year; from its June record high, it’s now down more than 20%, officially entering technical bear market territory.

However, behind the wild stock price swings, the simultaneous acceleration of lead times and rising prices suggests a deeper structural issue is emerging—“Chipflation.”

Julia Hermann, Global Market Strategist at New York Life Investments, recently warned that “Chipflation”—the surge in prices for AI-related logic and memory chips—will be the next headwind to test the resilience of AI trading.

In an interview, she noted: “Major cloud service providers now face a dilemma: on one hand, costs continue to rise—higher chip prices on top of climbing energy and utility bills; on the other hand, it may take years before investment returns materialize. We believe this environment will truly test the market's faith—so long as investors remain convinced of AI trades' long-term potential, they may tolerate short-term volatility and a slower pace of monetization.”

Asian markets are starting to provide evidence for these concerns. Hermann points out that one of the best indicators of memory chip inflation is the Korea DRAM export price index. In past cycles, memory chip year-on-year price growth peaked around 100%, but now, Korea's DRAM prices are up an astonishing 370% year-on-year.

In her view, while soaring chip prices are a sign of robust demand, they’re also a double-edged sword—persistently high prices will drive up the cost of building AI infrastructure, which in turn may suppress or even end the current wave of AI capex. Therefore, she is currently focusing on "quality" within the AI supply chain—defined as strong profitability, low-to-moderate earnings volatility, and ample interest coverage.

Choi Tae-won, chairman of South Korean memory giant SK Hynix, also recently warned about chipflation, bluntly stating that current high prices in the memory market are not normal. He expects that next year (2027), global semiconductor demand will expand substantially, with AI-related demand rising 60% to 100% over this year and overall semiconductor demand increasing at least 50% to 60%. However, next year's incremental supply will be “almost zero,” likely causing the supply-demand gap to widen further.

Addressing concerns that expanded production could prematurely end the current "supercycle," Choi Tae-won responded intriguingly: current chip prices are already abnormally high and should come down. If prices keep rising, further intensifying “chipflation,” the semiconductor industry will ultimately suffer. However, he made it clear that increasing supply and pushing prices lower doesn’t mean businesses can’t be profitable.

With supply-demand continuing to tighten and markets experiencing dramatic swings, the next phase for the semiconductor industry has become one of the most closely watched focal points in global capital markets.

0
0

Disclaimer: The content of this article solely reflects the author's opinion and does not represent the platform in any capacity. This article is not intended to serve as a reference for making investment decisions.

Understand the market, then trade.
Bitget offers one-stop trading for cryptocurrencies, stocks, and gold.
Trade now!

You may also like

Nvidia invests $1.5 billion in SB Energy, securing 8GW of AI computing power, with OpenAI set to become the sole tenant

Nvidia is extending its competitive edge from GPUs and servers to land, electricity, and construction for AI data centers. The company is partnering with SB Energy to develop an 8GW hyperscale AI data center in Ohio and is investing $1.5 billion to provide credit support for infrastructure financing, with OpenAI as the sole tenant. This move signals Nvidia’s shift from being solely a chip supplier to becoming an orchestrator of AI infrastructure, proactively securing LPS resources to meet the demand for multiple future generations of GPUs.

华尔街见闻2026/08/17 13:56