Tendrils is rethinking fundamental processor design to bypass historical constraints from the 1970s, including backwards compatibility, complex cache coherency protocols, and elaborate reorder buffers. To make its parallel architecture viable, the company is building an open-source programming language called Vine tailored specifically for parallel-first environments. The approach reflects the firm's core principle that future hardware requires ground-up co-engineering with software. Benchmark led the round, with Jack Altman taking a seat on the startup's board.
Key points
Tendrils intends to eliminate computing bottlenecks by discarding sequential chip designs…: Tendrils intends to eliminate computing bottlenecks by discarding sequential chip designs in favor of thousands of parallel cores.
The company seeks to abandon legacy microprocessor constraints such as backwards compatib…: The company seeks to abandon legacy microprocessor constraints such as backwards compatibility and complex cache coherency protocols.
Success relies on tightly coupling novel chip hardware with native parallel software engi…: Success relies on tightly coupling novel chip hardware with native parallel software engineering rather than adapting legacy stacks.
Sources checked by our analysis pipeline: 1
As of today, Tendrils has completed a $30M round led by Benchmark to rebuild general-purpose computing.
Contemporary CPUs are made under the constraint of backwards compatibility with sequential, single-core chips. Accumulated layers of historical designs are locking engineers into decisions dating back to the 1970s. As a result, all existing CPUs are bottlenecked by complex cache coherency protocols and elaborate reorder buffers. CPU scaling is sabotaged from the start. At Tendrils, we believe that the future of general-purpose computing lies beyond legacy architectures for legacy software.
Tendrils is answering the question: what if we redo the CPU? Instead of starting from a sequential design, we’re building our processor to have thousands of cores operating in parallel. This requires moving beyond the notion of shared mutable memory and linear instruction sequences. To make this possible, we’re working in tandem on the open-source language Vine, which has been designed from the beginning to target a new wave of parallel-first architectures.
We’re delighted to announce that Tendrils has raised $30 million to support this vision. This round is led by Benchmark, with significant…
Recent industry evaluations point to a notable shift in technology employment trends amid the rapid expansion of artificial intelligence. While product managers are emerging as the primary beneficiaries of this transformation, hiring interest for engineering talent has started leveling off. These observations suggest evolving priorities for technology organizations adjusting their operational strategies.
The social network is expanding its creator toolset by allowing users to trim longer videos into shorter snippets paired with custom commentary. This update supports a broader strategy to rival platforms like YouTube by encouraging short-form content circulation while keeping traffic contained on the site. Posts include an attribution link to drive audiences back to the initial video source.
Automatic Super Resolution has moved past its initial docked testing phase to provide AI upscaling on portable hardware and select Copilot Plus PCs. Operating directly on battery power at 1080p, the feature targets noticeable visual enhancements, particularly for legacy gaming catalogs. The rollout also removes earlier framerate constraints, paving the way for higher refresh rate portable gaming.