Microchip Technology and Micron are demonstrating a PCIe 6.0-based storage system at FMS 2026, the annual memory and storage industry conference being held in Santa Clara, California, from Aug. 4 to Aug. 6.

The demonstration combines Microchip's Switchtec PCIe 6.0 switch with Micron's 9650 NVMe solid-state drive (SSD).

PCIe switches distribute PCIe connectivity from a central processing unit (CPU) to multiple devices, including SSDs and graphics processing units (GPUs), enabling high-bandwidth communication across data center systems.

According to Microchip, the demonstration connects the entire data path—from the host processor to the SSD—using PCIe 6.0 technology to illustrate the real-world performance benefits of the new standard.

PCIe 6.0 supports transfer rates of 64 gigatransfers per second (GT/s) per lane, doubling the bandwidth available under PCIe 5.0.

Microchip's sixth-generation Switchtec PCIe switch is manufactured using a 3-nanometer process technology.

The device includes multicast functionality, allowing data to be distributed simultaneously to multiple endpoints, as well as secure boot capabilities based on post-quantum cryptography.

The switch also supports disaggregated infrastructure architectures, enabling storage resources to be pooled and dynamically allocated across servers as needed.

"Improving data center performance cannot be achieved through any single component alone," said Brian McCarson, Vice President of the Data Center Solutions Business Unit at Microchip.

"By combining the strengths of switching and storage technologies, we aim to demonstrate a scalable architecture for next-generation AI platforms," he added.

The PCIe 6.0 ecosystem is beginning to move into commercial deployment following AMD's recent launch of Venice, the first server CPU to support the new interface standard.

At FMS 2026, several companies are showcasing PCIe 6.0 products, including Samsung Electronics and Kioxia with PCIe 6.0 SSDs, as well as South Korean fabless semiconductor company FADU, which is presenting a PCIe 6.0 SSD controller.

Dutch semiconductor company NXP Semiconductors will supply ultra-wideband (UWB) chips to BMW for use in next-generation vehicle access and occupant detection systems.

NXP announced on Aug. 4 that its Trimension NCJ29D6 automotive UWB chipset will be integrated into selected 2026 BMW Group vehicle models.

BMW plans to use the technology to support its Digital Key Plus feature, which allows drivers to use smartphones and smartwatches as vehicle keys. The system will also enable occupant detection functions capable of identifying people or pets left inside a vehicle.

UWB is a wireless communication technology that can measure distances between devices with accuracy down to a few centimeters. In the automotive sector, it has primarily been used for digital keys and secure vehicle access control.

NXP said the NCJ29D6 is the first automotive UWB chip to integrate radar functionality.

By transmitting radio signals inside the vehicle and analyzing reflected waves, the chip can detect subtle movements such as human breathing, allowing the system to determine whether occupants remain inside the cabin.

According to NXP, the integrated architecture enables automakers to add multiple UWB-based features using the same hardware platform.

"Automakers can add UWB-based functions such as kick-motion recognition, intrusion detection and automated charging on the same hardware platform," said Markus Staeblein, Senior Vice President of Secure Car Access at NXP.

AMD expects its data center business to more than double in 2027 as demand for artificial intelligence infrastructure continues to accelerate, according to comments made during the company's second-quarter earnings conference call.

"We expect data center revenue in 2027 to more than double year over year," AMD Chair and Chief Executive Officer Lisa Su said on Aug. 4.

AMD's data center segment consists of its server CPU business and its GPU-based AI accelerator business.

Su said server CPU revenue is expected to increase more than 80% in the second half of 2026 compared with the same period last year. She added that server CPU sales are projected to grow by more than 70% again in 2027.

During the question-and-answer session, Su also said AMD's data center AI business is expected to continue growing well above 100%, with AI accelerators expanding faster than server CPUs and driving overall segment growth.

AMD has recently secured a series of major AI data center customers. Following agreements with OpenAI and Meta last month, the company also signed a strategic partnership with Anthropic.

Under the agreement, Anthropic plans to deploy AMD's next-generation AI rack system, Helios, at a scale of up to 2 gigawatts. The first 1-gigawatt deployment is scheduled to begin in the first half of 2027. Microsoft has also committed to deploying Helios within its Azure cloud platform.

Su said initial demand for Helios has exceeded expectations.

"Demand for Helios is stronger than we originally anticipated," she said. "We plan to begin shipments in the third quarter and expand deliveries through the fourth quarter and into next year."

The company also reiterated that supply constraints remain a challenge for its server CPU business.

"The server CPU supply chain remains tight. It has been tight throughout the first half of the year," Su said. "A significant portion of demand was not included in our forecasts."

She had previously acknowledged supply shortages during AMD's first-quarter earnings call in May.

AMD said it is increasing capacity across its supply chain, including wafers, advanced packaging and substrates. The company expects supply conditions to improve in 2027.

AMD now estimates the server CPU market will reach approximately $220 billion by 2030, implying annual growth of more than 50%.

The forecast marks a significant increase from previous projections. During its Analyst Day event in November 2025, AMD estimated annual growth of 18% for the server CPU market. In its first-quarter earnings call in May, the company raised its 2030 market estimate to $120 billion, implying annual growth of 35%.

AMD cited rising demand driven by the adoption of agentic AI as a key reason for the latest upward revision. While smaller than the company's projected $1.4 trillion AI accelerator market, AMD said the server CPU segment is expected to grow faster, at more than 45% annually.

The company has also begun production of its next-generation sixth-generation EPYC processor, codenamed Venice, which is manufactured using a 2-nanometer process technology.

According to Su, demand for Venice is the strongest of any EPYC generation to date.

AMD reported second-quarter revenue of $11.54 billion, up 50% year over year and a company record.

Data center revenue surged 107% to $6.7 billion, increasing its share of total revenue from 42% a year ago to 58%.

For the third quarter, AMD forecast revenue of approximately $13 billion, plus or minus $300 million. The outlook represents 41% year-over-year growth and exceeds the consensus analyst estimate of $12.52 billion.

"We are still in the early stages of a multi-year AI adoption cycle," Su said.

Despite the strong results, AMD shares fell more than 8% in after-hours trading after rising 7% during regular trading hours. While the company's revenue outlook exceeded Wall Street expectations, some investors viewed its AI growth projections as falling short of elevated market expectations.