Thursday, April 4, 2024

#OFC24: Transforming Cloud Interconnects with 800ZR/ZR+

At #OFC24, Lumentum announced significant performance enhancements to its 800ZR+ transceivers, enabling 800ZR+ modules to serve a wider range of applications, including expanded use in metro and regional networks through the ability to interface directly into routers. Additionally, the enhanced performance modes now make 400ZR+ and 600ZR+ operation suitable for true long-haul applications, with OSNR and dispersion tolerance enabling reaches exceeding 2000 km at 400 Gbps.  Marc Stiller, VP of Product Line Management from Lumentum, provides an in-depth explanation:

- Lumentum's newest modules, the 800G ZR+ and 400G high-power ZR+ modules, capable of running over a 400 km optical line system. These modules can plug into various routers and transmit over 2,000 km.

- He demonstrates a live traffic run at 800G, showcasing the simultaneous operation of two 800G channels and three 400G channels through an amplified 400 km length.

- Impressive results of the modules' performance, highlighting their ability to operate over 6,800 Pico seconds per nanometer of dispersion in the fiber 400 km, and their low error rates, indicating the potential for further extension of the links.

Want to be involved our video series? Contact


#OFC24: Scaling Silicon Photonics to Meet the AI Challenge

Check out OFC Conference and Exposition 2024 videos here:

Are we underestimating the market size and manufacturing requirements of optical modules needed to AI? Adam Carter, CEO from OpenLight, discusses:

- The increasing demand for high-speed modules in AI will surpass previous industry standards set by data centers, presenting unique challenges for manufacturers.

- To meet this demand, manufacturers will need to scale up and adopt production volumes similar to those seen in the Silicon industry for electronics today.

- OpenLight's strategic partnership with Jabil, a high-volume manufacturer, is a key step towards meeting these new demands. As part of this partnership, OpenLight will provide Jabil with taste test capability and on-wafer burning for their heterogeneous integrated silicon wafers.

Want to be involved our video series? Contact


##OFC24: Intelligent Re-timers for Linear Pluggable Optics

TeraSignal, a start-up based in Irvine, California, introduced an  intelligent 400G (4x100G) PAM4 modulator driver with digital link training and link monitoring for 800G linear pluggable optical (LPO) modules. The novel modulator driver addresses the limitations of LPO modules, which unlike DSP-based re-timers, have lacked the capabilities for automatic adaptation, digital link monitoring and link training, making them challenging to integrate, interoperate, and deploy efficiently. 

Armond Hairapetian, Founder and CTO from Terasignal, explains:

- The introduction of an intelligent retiree for LPO applications, which eliminates the need for a DSP inside the module

- The transition from Silicon germanium design to CMOS, which not only reduces power but also introduces diagnostic features in the chip, allowing end users to assess link quality and extract vital diagnostic information.

- The implementation of a digital subsampling scope on chip, which enables the capture of the Eye diagram, SNR, and BR of the link, providing end users with precise information about the link's performance and the ability to adjust the transmitter as needed.

Want to be involved our video series? Contact

#OFC24: Interoperability with LPO & LRO at 800G and 1.6T

Linear receive optics (LRO) and linear drive pluggable optics (LPO) were hot topics at #OFC24.  However, how to overcome the interoperability issue in LPO and LRO systems? New Photonics is introducing a second generation photonic integrated circuit (PIC) with integrated optical equalizer capable of supporting 800G and 1.6T modules. 

Professor Yosef Ben Ezra, CTO & Founder from New Photonics explains:

- The novel solution based on Optical Equalization of channels, which allows for interoperability between different vendors.

- The programmable nature of the Optical equalizer, which can be adjusted to the real Channel, ensuring no latency is introduced into the system.

- The integration of lasers, modulators, and Optical equalizers on the photonic integrated circuit, which dramatically improves the performance of LPO and LRO systems.

Want to be involved our video series? Contact


Lambda secures $500M for its GPU cloud

Lambda, which offers a cloud GPU service powered by NVIDIA, secured a special purpose financing vehicle of up to $500 million to fund the expansion of its on-demand cloud offering.

The special asset-based loan is secured by the GPUs and supported by their cash flow generation. 

The financing was led by Macquarie Group  with participation from Industrial Development Funding, and follows Lambda’s $320 million Series C funding round in February 2024.

“This first-of-its-kind financing vehicle unlocks a new way to fund the deployment of tens of thousands of NVIDIA GPUs,” said Mitesh Agrawal, Lambda COO. “Lambda’s mission is to continue to build the best AI compute platform in the world, and this structure lets our AI startup customers build the next generation of iconic companies on top of our cloud without the barrier of expensive multi-year contracts.”

The company cites explosive demand for its Lambda Cloud, which has more than 100,000 customer sign-ups. This financing supports NVIDIA GPU deployments for AI developers training, fine-tuning, and inferencing generative AI models in Lambda Cloud. Lambda also supports NVIDIA AI Enterprise to deliver a complete stack for generative AI developers.

“As demand for computing picks up pace alongside the widespread integration of LLMs and generative AI, we’re happy to support Lambda as they deliver what AI engineers need,” said Don Trent, Senior Managing Director in Macquarie Group’s Commodities and Global Markets business. “We’re proud to partner with Lambda on such a pioneering effort and support the buildout of GPU cloud infrastructure that will transform industries and businesses alike.”

Vertiv launches 360AI suite of power and cooling systems for data centers

Vertiv announced a new 360AI portfolio of high-density data center infrastructure solutions to support the higher cooling and power requirements of the accelerated computing IT stack. 

The new Vertiv 360AI suite is supports the power and cooling needs of artificial intelligence (AI) workloads of varying scales. This suite encompasses a range of products and services including power supplies, cooling systems, enclosures, lifecycle services, and digital management tools. 

The company says its suite is versatile, catering to a wide range of applications from AI test pilots and edge inferencing to large-scale AI data centers. The initial offerings are capable of handling up to 100kW per rack. Vertiv 360AI employs a variety of cooling strategies such as air-to-liquid, liquid-to-air, liquid-to-liquid, and liquid-to-refrigerant cooling to ensure optimal performance across different scenarios.

An important feature of Vertiv 360AI is its modular approach, which includes prefabricated IT, cooling, and power modules. These can be integrated and scaled according to specific needs, providing a flexible foundation for both current and future infrastructure demands. For environments lacking chilled water facilities, Vertiv 360AI solutions can incorporate indoor split chillers, offering a refrigerant-based liquid cooling option.

In addition to AI-specific solutions, Vertiv also presents an extensive selection of cooling and power technologies suited for high-performance computing (HPC) environments. This includes rear door heat exchangers, direct-to-chip liquid cooling, coolant distribution units (CDUs), free-cooling chillers, and modular data center solutions based on chilled water systems. Power infrastructure components such as uninterruptible power supplies (UPS) with lithium-ion batteries, busbar systems, switchgear, and battery energy storage systems (BESS) are also available to support microgrid applications, ensuring a comprehensive ecosystem for AI and HPC workloads.

Airspan enters Chapter 11, raises new funding, and delists

Airspan Networks confirmed that its stock has been delisted (NYSE American: MIMO). The company does not intend to  appeal NYSE Regulation’s determination.

Last week, Airspan and its U.S. subsidiaries  filed voluntary prepackaged Chapter 11 proceedings in the United States Bankruptcy Court for the District of Delaware. The company has entered into a Restructuring Support Agreement with certain funds managed by Fortress Investment Group and several of its other key financial stakeholders that will provide up to $95 million of new equity financing and the elimination of all the company’s existing funded debt.

“This support agreement is the culmination of a strategic review process, and we believe it is the best path forward for Airspan to continue providing exceptional services and products to our customers worldwide,” said Glenn Laxdal, President and Chief Executive Officer at Airspan. “By strengthening the Company financially with new capital and a debt-free balance sheet, we will be better positioned to execute our plan to capitalize on the significant growth opportunities across our public and private network markets. We appreciate the support and engagement of all of our stakeholders as we build Airspan for the future.”