In The News

Well, Google’s TPUs have gained immense market spotlight, but it appears that the firm might face a ’legal’ trouble, which could force them to change the name of their AI chips.

Trademark Protection Action: Tachyum, which applied for the TPU trademark in 2015 and secured rights in 2020, is demanding that Google cease using TPU to describe its products, indicating its legal claims to the trademark to maintain its competitive position in the AI market.

Tachyum has announced its new and open-source TDIMM memory standard, offering big uplifts in bandwidth & high capacity per module.

LAS VEGAS–(BUSINESS WIRE)–#Linux—Tachyum® today announced details about how its TDIMM™ technology is bringing the future of AI and computing, enabling AI models with parameters many orders of magnitude greater than those of any existing solution at a fraction of the cost. TDIMM is key to reducing the estimated cost of OpenAI data center $3 trillion and 250,000 megawatts of power to $27 billion and 540 megawatts.

The global AI infrastructure market is at a crossroads. As demand for large-scale AI models surges, the cost and energy barriers to deployment have become existential challenges for enterprises and governments alike. Enter Tachyum, a company whose Prodigy Universal Processor and proprietary TDIMM (Tachyum DDR5 DIMM) are positioning themselves as a disruptive force. By combining unprecedented memory bandwidth, energy efficiency, and cost reductions, Tachyum claims to offer a solution that could democratize access to next-generation AI. But how credible are these assertions? Let’s dissect the technical and economic claims underpinning this bold vision.

Tachyum(R) today announced details about how its TDIMM(TM) technology is bringing the future of AI and computing, enabling AI models with parameters many orders of magnitude greater than those of any existing solution at a fraction of the cost. TDIMM is key to reducing the estimated cost of OpenAI data center $3 trillion and 250,000 megawatts of power to $27 billion and 540 megawatts.

Tachyum today announced details about how its TDIMM is bringing the future of AI and computing, enabling AI models with parameters many orders of magnitude greater than those of any existing solution at a fraction of the cost. TDIMM is key to reducing the estimated cost of OpenAI data center $3 trillion and 250,000 megawatts of power to $27 billion and 540 megawatts.

Tachyum Prodigy 2nm Processor Intro BProdigy Ultimate provides up to 21.3x higher AI rack performance than Nvidia Rubin Ultra NVL576. Prodigy Premium provides up to 25.8x higher AI rack performance than Vera Rubin 144. Technical details of the 2nm Prodigy, the 1st ever chip to exceed 1,000 PFLOPs on inference, will be published within a week. Nvidia Rubin delivers 50 PFLOPs.

Tachyum® today announced details regarding how performance improvements are achieved for its 2nm Prodigy® Universal Processor. It will enable AI models with parameters many orders of magnitude more than those of any existing solution at a fraction of the cost.

In a major milestone for semiconductor innovation, Tachyum Inc. has officially unveiled its 2 nm Prodigy Universal Processor, designed to deliver breakthrough performance in artificial intelligence.

US-based technology company Tachyum has developed the Universal Processor, the 2nm Prodigy Ultimate, that combines the functions of a CPU, GPGPU and TPU into a single homogeneous processor architecture that 25.8x higher AI rack performance and 10x lower power than competing products. The processor delivers breakthrough performance and efficiency for a wide range of applications, including Hyperscale, HPC and AI data centres.

Tachyum has officially revealed detailed specifications for its upcoming Prodigy Universal Processor. This new 2nm chip in its “Prodigy Ultimate” variant will allow Tachyum to provide “21.3x higher AI rack performance than Nvidia Rubin Ultra NVL576”. Their “Prodigy Premium” model should also provide “up to 25.8x higher AI rack performance than Vera Rubin 144”.

Tachyum has unveiled its next-generation Prodigy Universal Processor, built on a 2nm process node, claiming unprecedented performance for artificial intelligence workloads and high-performance computing.

Tachyum has announced the release of its groundbreaking 2nm Prodigy chips, promising up to 1024 cores and cutting-edge DDR5 memory support. These chips aim to challenge NVIDIA’s Rubin Ultra in the competitive computing landscape.

Tachyum has unveiled its 2nm Prodigy chips, which aim to offer up to 1024 cores and support super-fast DDR5 memory, and will compete against NVIDIA’s Rubin Ultra.

Prodigy Ultimate provides up to 21.3x higher AI rack performance than Nvidia Rubin Ultra NVL576. Prodigy Premium provides up to 25.8x higher AI rack performance than Vera Rubin 144. Technical details of the 2nm Prodigy, the first ever chip to exceed 1,000 PFLOPs on inference, will be published within a week. Nvidia Rubin delivers 50 PFLOPs.

Tachyum, the Las Vegas processor startup with a Universal Processor integrating CPU, GPU and TPU, claims that its 2nm Prodigy Ultimate processor provides up to 21.3x higher AI rack performance than Nvidia Rubin Ultra NVL576 delivering 1,000 PFLOPs on inference compared to Rubin’s 50 PFLOPs.

Tachyum today announced details and specifications for its 2 nm Prodigy Universal Processor, which will enable AI models with parameters many orders of magnitude larger than those of any existing solution at a fraction of the cost.

As Tachyum continues towards tape-out for Prodigy, the company has chosen to establish manufacturing, testing, assembly, ODM, logistics and support facilities in Taiwan to ensure worldwide shipments of the completed product without the need for intermediaries.

The biggest news is that Tachyum’s Prodigy processor will adopt a multi-chiplet design and each compute chiplet within that system-in-package (SiP) will feature 256 universal cores.

With a total of more than $300 million invested over three funding rounds, Tachyum is poised to revolutionize AI data centers by solving the most challenging barriers to AI and expects that Series C will propel the company to an IPO, potentially as early as 2027.

Tachyum’s AI team has demonstrated that a foundation model fine-tuned on a task-specific dataset in FP4 data type, which represents value using 4-bit floating-point format rather than standard FP32 or BF16, achieves parity with traditional FP32 training baselines.

Tachyum’s engineers ported Kprobes (Kernel Probes), which plays an important role in the eBPF JIT technology and serves as trigger for eBPF subroutines.

Tachyum is at the forefront, pushing the boundaries with its Prodigy Universal Processor which promises unparalleled versatility across various applications. A recent development has taken center stage with the successful port of the eBPF Just-In-Time (JIT) compiler to their software emulation platform.

The white paper explains that MoE can achieve performance comparable to dense models while using approximately four times less computing power and memory bandwidth, though memory capacity requirements increase by about four times.