A $90 million funding round announced on 18 November 2026 puts a San Francisco company called Linq at the centre of a shifting bet on how people will interact with software. Andreessen Horowitz led the Series B, and Ben Horowitz joined the board. Linq's products let AI agents send messages, make phone calls and operate software through a computer interface — the ordinary tools people use to get things done, repackaged as infrastructure for machines.

The company says it has processed more than one billion messages and counts DoorDash among its customers. In one example highlighted in the announcement, a DoorDash customer can find food, build a cart and check out inside a single Apple Messages conversation, with an agent handling the steps that once required several screens. Linq is also an approved Apple Messaging Service Provider, a status that matters in a channel where trust and deliverability are hard to replace.

The round speaks to a broader change in venture thinking. Instead of funding another chatbot, investors are backing the plumbing that lets agents reach users and external services reliably. This article explains the financing, Linq's unusual path from digital business cards to agent infrastructure, the products it is building, and the technical and security problems that remain unresolved.

Ninety million dollars for agent communications

Linq announced on 18 November that it had raised $90 million in new funding led by Andreessen Horowitz. The company is based in San Francisco and describes itself as an infrastructure platform for AI agents. The release was distributed through PR Newswire and reproduced by Yahoo Finance, giving a clear primary record of the terms.

Secondary coverage from TechStartups characterised the financing as a Series B and reported that total capital raised now exceeds $120 million. CapTables reported that the round valued Linq at $1 billion. That valuation is not stated in the company's own release, so it should be treated as reported rather than confirmed.

The investment comes at a moment when AI agents are moving from demonstration to deployment. Andreessen Horowitz's own note on the deal said 2026 was the year agents crossed into the mainstream, accounting for more than half of internet traffic. That figure is drawn from Cloudflare's network rather than a universal measurement, but it captures the scale of the shift investors see.

Linq is not building the models behind those agents. It is building the channels and tools they use to communicate and act. That distinction explains why the round is structured around infrastructure rather than a consumer product.

The round and the board seat

Andreessen Horowitz led the financing, with continued participation from TQ Ventures, Mucker Capital and Abstract. Anthos and Left Lane Capital joined as new investors. Ben Horowitz, co-founder and general partner at Andreessen Horowitz, will join Linq's board of directors.

A board seat is more than a ceremonial detail. It signals that the lead investor intends to be involved in the company's next stage, particularly as it expands into voice calling and computer use — areas where product decisions, partnerships and security standards are still being defined. Horowitz's public comment focused on the core idea: agents will communicate and get work done much like people do, and Linq gives them infrastructure to do both at scale.

Elliott Potter, Linq's co-founder and chief executive, confirmed the round and the cumulative funding total in a post on X, according to TechStartups. He said the company had raised more than $120 million to date. The new capital will support continued investment in messaging and expansion of the product line through software, hardware infrastructure and strategic partnerships.

The investor list combines earlier backers with new specialist funds. For a company that has reinvented itself several times, continuity from TQ Ventures, Mucker Capital and Abstract is a useful signal. It suggests some investors stayed through the pivots that preceded the current strategy.

From digital business cards to agent infrastructure

Andreessen Horowitz's investment note provides the most detailed account of Linq's history. The firm first met Potter in 2020, on a video call during the pandemic, when Linq was building digital business cards in Birmingham, Alabama. The founders later moved the company to San Francisco to be closer to customers.

The product evolved from digital business cards to software for sales teams, then to AI sales agents, and finally to communications infrastructure. According to a16z, the last transition came when customers began asking to use the messaging technology underneath the application for their own purposes. The founders recognised that the infrastructure itself was the opportunity and rebuilt the company around it.

This history matters because it explains the company's positioning. Linq is not a newcomer to messaging infrastructure; it spent years building the underlying system for its own applications before selling it to others. a16z described the founders as consistently underestimated while they were based in Alabama and said Potter had written publicly about nearly running out of money several times.

The three co-founders are Potter, chief technology officer Patrick Sullivan and president Jared Mattsson. Their journey from a small Alabama startup to a $90 million round is unusual in a sector where companies often pivot quickly and attract capital at the first sign of traction.

Why messaging came first

Messaging was Linq's first infrastructure product and remains central. The company says it has processed more than one billion messages. Its customers include DoorDash, Town and Tomo, according to the press release; TechStartups listed DoorDash, Town and Tab. The precise customer roster varies between accounts, but DoorDash is consistent.

The appeal of messaging is practical. People already use text threads to coordinate almost everything, and an agent that can operate inside a familiar conversation does not require a new app or a new habit. Aryan Shah, head of AI research at DoorDash, said in the release that Linq gives the company messaging infrastructure to bring ordering into a familiar conversation, so customers can move from deciding what they want to placing an order in one thread.

Linq's status as an Apple Messaging Service Provider gives it a formal role in that channel. It supports Apple Messages for Business, along with RCS, WhatsApp and other platforms. The company's announcement cited Cloudflare's finding that more than half of internet traffic is no longer human and that AI agent requests rose more than 1,700% in a year.

Messaging also creates an audit trail. For businesses deploying agents that make bookings or purchases, a conversation record can help with customer support, dispute resolution and compliance. The press release does not make that argument explicitly, but it follows from the product's design.

Beyond text: calling and computer use

The new funding is not only about messaging. Linq has introduced HD Calling for Agents, which gives agents the ability to make phone calls. That capability could be used to schedule appointments, contact businesses or retrieve information that is not readily available online.

The company is also building what it calls computer use: access to physical computing infrastructure that lets agents interact with software and complete tasks through computer interfaces. Kernel, a company focused on browser automation, is among the early customers for that capability, according to TechStartups and Blockster.

These products represent a significant expansion in scope. Sending a message is a relatively contained operation. Allowing an agent to use a browser, fill forms or initiate transactions introduces questions about permissions, identity and error handling. The company's own language is cautious: Potter said agents should be as easy to reach as a friend and capable enough to finish the job.

The strategic logic is clear. If agents become a primary interface between customers and businesses, the company that supplies messaging, voice and computer access could sit underneath many applications. Linq wants to be that layer rather than one of the applications built on top of it.

The billion-message proof point

More than one billion processed messages is Linq's most concrete operating metric. It is a measure of platform activity, not revenue or profitability, and the company has not published financial figures. Still, it gives some evidence that the infrastructure works at scale.

Volume matters in messaging infrastructure because reliability is difficult to retrofit. Deliverability, throughput, encoding, channel rules and integration with enterprise systems all have to work before a customer will trust an agent with a transaction. A billion messages suggests Linq has already passed a meaningful operational threshold.

The metric also helps explain investor interest. In a market full of AI startups with impressive demonstrations but limited usage, a company that can point to a large, repeating workload stands out. a16z said many of the fastest-growing agent companies already use Linq to connect their agents directly to customers.

There are limits to what the number proves. Messages can range from routine notifications to complex transactional threads, and the release does not break them down. Nor does it say how many customers generate the volume. The figure is best read as evidence of scale, not as a full picture of the business.

DoorDash and the Apple Messages channel

The DoorDash example is the most specific customer story in the public record. Through Apple Messages for Business, a customer can find food, build a shopping cart and complete checkout without leaving the conversation. The press release quotes Shah saying the approach keeps the ordering experience inside a familiar channel.

This is a different model from a chatbot embedded in a company's own app. Instead of drawing the user into a proprietary interface, the agent meets them where they already are. For large businesses with high-frequency customer interactions, that can reduce friction and support costs. For Linq, it provides a reference customer with recognizable consumer reach.

Apple's role is important. As an official Messaging Service Provider, Linq operates within Apple's rules for business messaging. That status creates a barrier to entry and gives enterprise customers confidence that the integration is legitimate. a16z emphasised the point in its investment note, calling Linq an approved provider.

The same model can extend to other channels. Linq says it supports RCS and WhatsApp today, with the ambition of helping businesses and agents communicate wherever users are. Whether the economics of each channel are comparable remains to be seen, but the product architecture appears designed for that expansion.

The agentic traffic thesis

Both Linq and its lead investor are betting on a structural change in internet usage. Cloudflare's 30 September 2026 post, titled "The Internet has a second audience", reported that daily requests from AI agents on its network grew by more than 1,700% over a year. It said that for the first time, more than half of internet traffic was not human, and that its network handled about 115 million HTTP requests per second, up from 63 million at the end of 2024.

Those numbers describe Cloudflare's network, not the whole internet, and the company itself notes that different categories of automated traffic should not be conflated. Even with that caveat, the direction is dramatic. Businesses that once optimised for human visitors now have to serve software acting on behalf of people.

Linq's products are aimed squarely at that audience. An agent booking a table, comparing insurance quotes or buying a dataset is a customer in Cloudflare's framing, even though no human clicks through the site. Linq's messaging and calling infrastructure is designed to let such agents complete the interaction.

The thesis has an important commercial implication. If agents become the primary way customers reach companies, the interface layer could be as valuable as the underlying model. a16z's investment note states this directly: companies will need to adapt so their customers can communicate with them through agents, and Linq is positioned as the platform for that adaptation.

Where the capital goes

The company's announcement says the funding will support continued investment in messaging and expansion of the product line through its own software, hardware infrastructure and strategic partnerships. That is broad language, but the priorities can be inferred from the products recently introduced.

Messaging will remain the base. Voice calling will require carrier relationships, audio quality controls and new compliance processes. Computer use will require computing infrastructure, isolation between agent sessions and careful handling of credentials. Each of those areas needs engineering and partnership work rather than model training.

The mention of hardware infrastructure is notable. Many AI infrastructure companies describe themselves as software businesses. Linq's willingness to invest in physical computing reflects its view that agents need real machines to interact with software, not just API calls to other services.

Geographic expansion is also likely, though it should not be assumed. The company moved from Birmingham to San Francisco to be closer to customers, and its investor base now spans established venture firms and new specialist funds. a16z noted that the founders spent their first six years in Alabama before relocating, which suggests they are willing to make structural changes when customer needs demand it. No specific international offices were announced in the release, so any such plan would be speculation rather than reported fact.

The messaging business also carries operational obligations that do not appear in a product description. An approved Apple Messaging Service Provider must meet platform requirements, and enterprise customers will expect service levels that match the systems agents are replacing. Those obligations are part of what the new capital is likely to fund, even if the announcement groups them under the broad heading of continued investment.

The reliability and security problem

The biggest obstacle facing Linq is not demand. It is trust. An agent that sends a message is relatively easy to control. An agent that completes a transaction, calls a business or operates a browser must be constrained by permissions and must behave predictably when something unexpected happens.

TechStartups highlighted reliability as the central challenge. Allowing an agent to interact with third-party services and make decisions within user-defined permissions introduces greater technical and security demands than simple message delivery. Those demands include authentication, authorisation, auditability and the ability to stop an agent before it completes an unintended action.

The computer-use product raises the stakes further. Browser automation has long been used for testing and data collection, but giving an AI agent access to a live session with payment or account privileges is a different category of risk. Kernel's involvement suggests Linq is starting with a specialist partner rather than building every capability in-house.

There is also the question of how much work agents can reliably complete. a16z's note describes agents planning trips, booking restaurants and scheduling meetings much like a coworker. Today's systems still require oversight for many such tasks. Linq's infrastructure can improve the interface, but it cannot by itself solve model reliability.

What Linq's round says about the AI stack

The funding is best understood as a bet on the layer between AI models and the services people use every day. Large model developers will continue to compete on intelligence. Cloud providers will compete on compute. Linq is competing on access — making sure an agent can reach a customer through a message or a call, then carry out a task through software.

That is a defensible position if agents become as common as websites or apps. It is less defensible if interaction stays concentrated inside a handful of dominant consumer platforms, or if businesses build their own agent channels. The next few years will determine which of those scenarios is closer to reality.

For now, the facts are limited but concrete. Linq raised $90 million on 18 November 2026 in a round led by Andreessen Horowitz, with Ben Horowitz joining the board. It says it has processed more than a billion messages, serves customers including DoorDash, and is expanding into voice and computer use. The capital gives it room to build. The difficult part will be proving that agents can be trusted with the work the company wants them to do.


Corrections and updates

Nexuswild welcomes factual corrections. Email contact@nexuswild.com with evidence and the article URL.