Volantis Raises $88 Million to Tackle AI’s Biggest Bottleneck: The Race to Make Artificial Intelligence Faster

Hookline: The next breakthrough in AI may not come from building a bigger model, but from rethinking how quickly its processors can access the information they need.

Artificial intelligence is advancing at an extraordinary pace, transforming software development, business operations and the way people interact with technology. Yet behind the increasingly capable AI models lies a persistent engineering challenge: moving information between processors and memory quickly enough to keep up with demand.

San Francisco-based semiconductor startup Volantis is taking on this challenge with an ambitious approach that uses light to improve communication between computing and memory components. On October 1, 2026, the company announced an $88 million Series A funding round, bringing its total funding to approximately $97 million. The investment will support the development and commercialization of its photonics-based AI inference architecture.

The funding round was co-led by investor Lachy Groom and Abstract Ventures, with participation from prominent technology investor John Doerr, VXI Capital, Triatomic and Susa Ventures. Angel investors, including Dwarkesh Patel, Naveen Rao and Sholto Douglas, also participated. Source: Reuters.

The announcement has drawn attention to an increasingly important segment of the AI industry: the hardware infrastructure that determines how efficiently advanced models can operate.

As businesses deploy larger AI models and increasingly sophisticated AI agents, the ability to process information quickly and economically is becoming a competitive advantage. Volantis believes that addressing memory limitations could help unlock the next phase of AI performance.

Volantis Raises $88 Million: Inside the Funding Round

Volantis announced its $88 million Series A on September 29, 2026, through a company funding announcement, with coverage from Reuters and other technology publications following on October 1 and in the days afterward.

According to the company, the latest investment brings its total funding to approximately $97 million, including an earlier $9 million seed round reported in June 2025. Source: Volantis.

The funding details are as follows:

• Company: Volantis
• Funding round: Series A
• Amount raised: $88 million
• Total funding reported: Approximately $97 million
• Round co-leads: Lachy Groom and Abstract Ventures
• Other participating investors: John Doerr, VXI Capital, Triatomic and Susa Ventures
• Angel investors: Dwarkesh Patel, Naveen Rao and Sholto Douglas
• Primary focus: Developing and commercializing photonics-based AI inference hardware

The investment is significant because it targets a highly specialized area of technology that could influence the performance of future AI systems.

While many AI startups focus on developing applications, software platforms or foundation models, Volantis is concentrating on the underlying hardware architecture. Its goal is to improve the movement of data between processing units and memory, potentially making advanced AI workloads faster and more cost-efficient.

What Is Volantis, and Why Does Its Technology Matter?

Volantis is a semiconductor startup developing a new architecture for AI inference.

Inference is the process through which a trained AI model generates an answer, writes code, analyses information or performs another task. It is a central component of everyday AI applications, from chatbots and coding assistants to enterprise automation platforms.

As AI adoption grows, inference systems must handle increasingly complex workloads. Larger models require substantial memory capacity, while high-speed applications need sufficient bandwidth to transfer information between memory and processors.

These requirements create a difficult engineering trade-off.

A system may have enough memory to accommodate a large model but struggle to feed information to the processor quickly enough. Alternatively, a system designed for very high bandwidth may have limitations in the amount of information it can store close to the computing hardware.

Volantis is attempting to address this challenge by developing an optical interconnect architecture that uses light to transfer data between computing and memory components.

Rather than relying exclusively on conventional electrical connections, its approach uses photonics to expand the potential capacity and bandwidth of an AI inference system.

The company’s ambition is to make it possible to run substantially larger models at higher speeds while reducing the cost of generating each token, the basic unit of text or other information produced by an AI model.

The AI Memory Bottleneck: A Challenge Facing the Industry

The growing complexity of AI models has placed greater pressure on the hardware supporting them.

Modern AI accelerators, including GPUs, depend on high-bandwidth memory to provide rapid access to model parameters and intermediate data. However, conventional electrical interconnects impose physical constraints on how computing chips and memory components can be connected.

According to Reuters, Volantis says its optical approach could enable up to 220 memory chips to be connected around a GPU, compared with the eight high-bandwidth memory chips described in the report for Nvidia’s current leading configurations.

This is a company-reported architectural goal, not evidence that a commercially deployed Volantis system has already achieved that configuration.

The underlying challenge is widely relevant to AI infrastructure. As models expand, memory capacity and data-transfer bandwidth can limit performance even when substantial processing power is available.

Volantis is targeting this limitation through a design intended to improve both the amount of memory available to a processor and the rate at which information can move between components.

If the company can demonstrate its proposed advantages in practical systems, its technology could offer an alternative approach to scaling AI inference hardware.

How Photonics Could Change AI Computing

One of the most distinctive elements of Volantis’s strategy is its use of photonics.

Photonics is the science and technology of generating, controlling and detecting light. Optical communication already plays an important role in telecommunications and data-centre networking, where light can transport information over distances and at bandwidths that are difficult to achieve with conventional electrical connections alone.

Volantis is applying optical technology to a different problem: improving communication between AI processors and memory.

The company’s architecture uses vertical-cavity surface-emitting lasers, commonly known as VCSELs. These compact laser devices are already used in commercial applications, including sensing and facial-recognition systems in consumer electronics.

According to Reuters, Volantis hopes that using this established technology could help reduce some of the supply-chain and manufacturing challenges associated with developing new semiconductor systems.

The company is also working on an integrated optical technology stack designed to deliver high bandwidth density without relying on conventional external lasers and traditional optical fibre in the same way as some existing approaches.

Its first planned system, called A-1, is being designed to support models exceeding 20 trillion parameters and target speeds of up to 10,000 tokens per second per user, according to the company’s announcement.

These are ambitious development targets, not independently verified performance results. Achieving them in production will depend on engineering execution, software compatibility, system integration and real-world testing.

Nevertheless, the approach illustrates how innovations in memory architecture and data transfer could become increasingly important as AI systems scale.

Why AI Agents Could Benefit From Faster Inference

The rise of AI agents provides an important commercial context for Volantis’s technology.

Traditional AI assistants primarily respond to prompts, generate information or help users complete individual tasks. AI agents are increasingly being developed to execute multistep workflows, interact with software tools, write and test code, and coordinate activities with limited human intervention.

These tasks can require repeated model inference calls. The speed of each interaction can influence the time needed to complete an entire workflow.

For example, a coding agent might inspect a codebase, identify an issue, generate a fix, run tests and revise its approach based on the results. Faster inference could help reduce the time required for such tasks, provided that the rest of the workflow can keep pace.

Volantis has highlighted this opportunity in its own funding announcement, describing a future in which a coding agent could complete work in seconds or minutes rather than taking much longer.

Such outcomes remain illustrative rather than guaranteed. End-to-end performance also depends on model quality, software execution, external tools and the complexity of the task.

Still, the potential market is significant. Faster, more cost-effective inference could benefit AI developers, cloud infrastructure providers and businesses deploying AI agents at scale.

The Team and Investors Behind Volantis

Volantis’s fundraising has also attracted attention because of the experience represented by its team and investors.

The company says its founding team includes semiconductor and photonics specialists with backgrounds at NVIDIA, AMD, Broadcom and Ayar Labs.

Members of the team have worked on advanced semiconductor packaging, optical communication systems and high-volume laser technology. These are relevant areas of expertise for a company attempting to introduce a new architecture into a technically demanding market.

The investor group also includes established technology investor John Doerr, alongside venture capital firms and individuals with experience in the technology ecosystem.

Such backing can provide more than financial resources. Experienced investors may help early-stage hardware companies build industry relationships, recruit specialist talent and navigate the long development cycles associated with semiconductor products.

However, funding and team credentials do not guarantee commercial success. Volantis will ultimately need to demonstrate that its technology can deliver meaningful benefits at a competitive cost.

What the Latest Volantis News Reveals About Its Direction

Publicly available news about Volantis over the past three months has been concentrated around its September–October 2026 Series A announcement and the technology behind the planned A-1 system.

The company’s September 29 funding announcement outlined its objective of using photonics to improve memory capacity and bandwidth. Reuters subsequently reported on October 1 that Volantis was pursuing a laser-based approach to connecting AI processors and memory chips. Specialist publication optics.org also examined the company’s claims and its use of integrated VCSEL technology on October 6.

Together, these reports provide a clearer picture of the company’s direction: Volantis is seeking to develop an alternative AI inference architecture rather than simply compete as another AI application provider.

The next developments worth watching will be progress on the A-1 system, demonstrations of real-world performance, manufacturing readiness and potential customer adoption.

As of the latest reports reviewed for this article, the company’s proposed performance figures should be treated as development targets rather than confirmed commercial results. No additional major funding round or independently verified commercial deployment was established by the sources reviewed.

Challenges Volantis Must Overcome

Despite the opportunity, bringing a new semiconductor architecture to market is a complex undertaking.

Manufacturing and reliability: The company must demonstrate that its optical components and integrated architecture can be produced consistently and operate reliably under demanding workloads.

Compatibility: AI infrastructure providers have invested heavily in established GPU, memory and software ecosystems. Volantis will need to show that its system can integrate effectively with relevant hardware and software.

Performance validation: The company must demonstrate that its proposed gains in memory capacity, bandwidth, speed and cost translate into measurable advantages for customers.

Competition: Semiconductor companies and infrastructure developers are investing heavily in high-bandwidth memory, advanced packaging, optical interconnects and other approaches to improving AI performance.

Commercial adoption: Even if the technology performs as intended, customers will need a compelling reason to adopt it, including clear economic benefits and confidence in long-term support.

The $88 million funding round provides Volantis with resources to pursue development and commercialization, but the company’s progress will depend on how effectively it addresses these challenges.

What Volantis’s Funding Means for the Future of AI

Volantis’s funding round reflects a broader shift in the AI investment landscape.

The initial wave of generative AI investment focused heavily on models and consumer-facing applications. As adoption expands, investors are also paying greater attention to the infrastructure required to operate those systems efficiently.

That includes semiconductors, memory, data-centre networking, power management and advanced computing architectures.

The next stage of AI development may depend not only on creating more capable models but also on making them practical to operate at scale.

Volantis is betting that photonics can help address one part of that equation by improving how processors access memory. If its approach succeeds, it could contribute to faster inference, larger workloads and potentially lower operating costs.

For enterprises, these improvements could make certain AI applications more responsive and economical. For AI developers, they could offer new options for running demanding models. For the semiconductor industry, they could help advance the broader exploration of optical technologies for computing infrastructure.

The opportunity is substantial, but it remains to be proven through working systems and commercial deployments.

Conclusion: Betting on the Infrastructure Behind AI

Volantis’s $88 million Series A funding round is an important development in the race to build more efficient AI infrastructure.

By focusing on the memory bottleneck, the startup is targeting a technical challenge that could become increasingly significant as models grow larger and AI agents take on more complex tasks.

Its photonics-based architecture offers a promising direction, supported by a team with semiconductor and optical technology experience and a group of high-profile investors.

The decisive test, however, will be execution. Volantis must demonstrate that its technology can achieve its performance targets, integrate into real-world systems and deliver economic benefits that justify adoption.

The larger lesson is that the future of artificial intelligence will be shaped by more than models alone. It will also depend on the chips, memory systems and connections that allow those models to work.

Volantis is making a bet that the next leap in AI performance could begin with a simple but fundamental change: using light to move information more effectively.

If the company can turn that vision into commercially viable technology, its progress could become an important story in the next chapter of AI infrastructure innovation.

written by 
Manisha