In early June 2026, Perplexity raised about $200 million at a valuation near $20 billion to scale Comet, its AI browser. The raise is a bet on owning the front door of the agent economy, the surface where AI agents start tasks and finish purchases.
- Comet launched in 2025 and went free worldwide that October. The June 2026 capital is fuel to scale it before better-funded rivals close the window.
- An AI browser inserts a machine reader between your customer and your store, so structured data and a checkout an agent can complete start to matter as much as design.
- The Amazon injunction from March 2026, now on appeal, will help decide whether agents shop your store freely or only where you let them in.
- The move for brands is low-cost insurance: clean feeds, agent-readable catalog, and AI referral tracking, not a panic pivot.
In early June 2026, Perplexity raised about $200 million at a valuation near $20 billion. The headline says the money is for Comet, its AI browser, but the browser is only the vehicle. What Perplexity is buying with that capital is a shot at owning the front door of the agent economy: the place a person opens when they want something done, and an agent goes off and does it, including buying.
There is a lot of noise around AI and commerce, and most of it is breathless, so it helps to separate what happened from what it might mean. The funding is real and recent. Its strategy was set in 2025, when Comet launched and then went free worldwide that October, so the new money scales an idea that already existed. If you build a DTC brand or a Shopify app, the useful question is whether this changes anything you do this quarter. For most of you it does, a little, and cheaply.
This is part of the same shift I keep coming back to in the agentic commerce primer for Shopify brands: the customer increasingly arrives through a machine, not a tab. Comet is one of the most aggressive attempts yet to be that machine. The raise is a signal that serious investors think the front door is worth fighting for, and that they think it is up for grabs.
The $200 million
is buying the spot
where shopping starts.
Today, when you want to shop, you open a browser or an app, you search, you compare, you decide, you check out. Each of those steps is a place where a company can sit and take a cut: Google on the search, Amazon on the marketplace, the brand on the checkout. An AI browser collapses all of those steps into one conversation with an agent, and whoever owns that conversation owns a new toll booth on the whole purchase.
That is the position Perplexity is paying for. Comet went free worldwide in October 2025 precisely to grab the surface area, because a free agent that millions of people use to start tasks is worth far more than a paid one that few do. The June 2026 capital scales that land grab. It is the same playbook every platform has run: get the default first, monetize the toll later. I have watched this pattern from the inside at Shopify, and the early-mover advantage in owning a default surface is real and durable.
"Whoever owns the conversation where a purchase starts owns a new toll booth on the whole journey. That is what the $200 million is actually buying."
The raise also tells you the race is crowded. Other companies see the same front door, and the capital is partly defensive: fuel to scale Comet before larger players close the window. For a brand, the point is that several well-funded companies are all trying to become the layer between you and your customer, and at least one of them probably will, whichever browser wins.
When an agent shops,
the front door
moves off your site.
For twenty years the front door to your store has been a search result, an ad, or a social post that lands a human on your product page. You optimized for that human: the photography, the copy, the reviews, the urgency. An AI browser changes who shows up first. Before a person sees your page, an agent reads it, weighs it against alternatives, and often decides without the human ever looking.
An agent reading your page ignores the hero image. What it needs is to read your price, your availability, your variants, your shipping, and your return policy in a structured form it can trust and act on. The brands that win the agent layer will be the ones whose data is clean and whose checkout an agent can complete, the same way the SEO winners of the last decade were the ones whose pages search engines could read cleanly. The payment side is already being built for that, with Visa and OpenAI putting agent tokens into checkout.
| Stage | Old front door (human) | New front door (agent) |
|---|---|---|
Discovery | Search, ads, social land a person on your page | Agent reads many sources, surfaces a shortlist |
Comparison | Human scans pages, reviews, prices | Agent compares structured data across stores |
What wins | Design, copy, reviews, urgency | Clean feeds, readable catalog, trust signals |
Checkout | Human completes the cart on your site | Agent completes it, where it is allowed to |
Who you optimize for | The shopper | The shopper and the machine reading for them |
The human still matters. People want to feel something about a brand, and that is still won with design and story, but a second reader now sits in front of the first, and that reader is literal, so the work is making your store legible to both. The mechanics of exposing your catalog to agents are in the catalog API and AI agents breakdown, and they are mostly plumbing.
The front door is
contested, and the
courts are in it now.
Before you assume agents will roam freely across the web buying things, look at what happened with Amazon. In March 2026, a federal judge granted Amazon a preliminary injunction blocking Comet's agent from accessing parts of Amazon's site, after Amazon argued the agent was accessing its store without authorization and disguising itself as a normal browser. The court found Amazon had strong evidence the access was unauthorized. Perplexity appealed in April, arguing that directing an agent to shop is no different from a person opening a browser and visiting the site.
That case is bigger than two companies. It is testing whether an agent acting at a user's direction counts as authorized access to a site, and every agent that shops on a user's behalf depends on the answer. I broke down the stakes of this in the post on the agent commerce legal precedent, because the outcome will shape what every merchant can and cannot do about agents at their own door. The same conflict showed up when ChatGPT's instant checkout got pulled back, and again in Amazon blocking OpenAI in the catalog war.
A fight between Amazon and Perplexity reaches your store too. The ruling will help set whether agents have a default right to shop any site, or whether merchants get to decide who they let in. As a DTC brand, you will likely want agents to be able to buy from you, which puts you on a different side of this than Amazon. Know where you stand before the precedent is set for you.
Do not assume the rules are settled. The legal and technical lines around agents are being drawn right now, in real cases, and they will determine whether the agent at your door is a welcome new channel or an unauthorized scraper. Most DTC brands will want to be on the welcoming side, because an agent that can buy from you is just another customer, and a high-intent one at that.
What an AI browser
actually changes for
a Shopify merchant.
Three things shift for a brand if agent-driven shopping keeps growing, and none of them require you to rebuild your store.
Your traffic source list grows a new line. Today you watch search, paid, email, and social. Soon you watch AI referrals as their own channel: people and agents arriving from Comet, ChatGPT, and the rest. If you cannot see that traffic, you cannot manage it. Tracking AI referral sources is the cheapest first move, and it tells you whether any of this is real for your specific brand yet.
Your product data becomes a sales surface. The feed you treat as back-office plumbing is what a machine reads to decide whether to recommend and buy you. A messy catalog, missing variants, or stale availability now costs you sales you never see, because the agent skipped you. Clean structured data, once an SEO nicety, is now a conversion lever.
Your reviews and trust signals get read literally. A human skims reviews for vibe. An agent parses rating, count, recency, and sentiment as inputs. The same trust signals that move human conversion now move machine selection, which means the work you already do on reviews and social proof pays twice. This connects to how brands defend their position as the channel mix shifts, which I covered in why enterprise keeps losing to DTC challengers.
Care about the trend,
do not bet the
company on the browser.
My read is deliberately unexciting. Agent-driven shopping is real and growing, and the front-door fight is worth watching closely. It is also early and contested, and full of companies that will not exist in three years. There are two ways to get it wrong: ignore it because it feels like hype, or pour resources into a single browser or protocol that may lose.
The right posture is to prepare the surface. Whether Comet, ChatGPT, or something not yet launched becomes the front door, the work to get your store ready is largely the same: clean data, an agent-completable checkout, and visibility into AI traffic. That is low-cost insurance against a high-probability shift, the kind of bet an operator should take.
I have lived through enough platform shifts to know the pattern. The brands that did well through the rise of mobile, social commerce, and marketplaces were rarely the ones who guessed the exact winner. They stayed ready and moved fast once a winner emerged, and I expect the same here. The agent layer is the same kind of structural shift I mapped in the 2026 Shopify ecosystem value map, and the durable move is to get positioned without betting on one outcome.
Four moves that cost
little and leave you
ready either way.
If you do nothing else after reading this, do these four things. They are cheap, they help your store today even if the agent layer stalls, and they leave you positioned if it does not.
Perplexity's raise is a useful marker. It tells you that smart money believes the front door of the agent economy is worth fighting for, and that it is still up for grabs. For a DTC brand or a Shopify merchant, the sensible response is to spend a little, get your store ready for the machine reader sitting in front of your customer, and watch the fight play out from a prepared position.
What brands ask me
about Comet and
AI shopping.
Comet is Perplexity's AI browser. Instead of you opening tabs and clicking through sites, you tell Comet what you want and an agent navigates, compares, and can act on your behalf, including shopping. It launched in 2025 and went free worldwide that October. The roughly $200 million Perplexity raised in June 2026 is fuel to scale Comet into the front door of the agent economy, the shift I cover in the agentic commerce primer.
The raise, reported in early June 2026 at a valuation near $20 billion, is a bet on owning the surface where AI agents start tasks, including buying things. Comet is the vehicle for it. Whoever controls that front door controls a slice of discovery and a cut of transactions. The capital is there to scale Comet before better-funded rivals close the window, the same land-grab pattern I map in the ecosystem value map.
It adds a new layer between your customer and your store. When an agent does the shopping, your product page is read by a machine first, not a person. Structured data, clean feeds, and a checkout an agent can complete start to matter as much as your design. The brands that get cited and transacted by agents will win share the way SEO winners won search. The plumbing is in the catalog API breakdown.
In March 2026 a federal judge granted Amazon a preliminary injunction blocking Comet's agent from accessing parts of Amazon, and Perplexity appealed. The case is testing whether an agent acting at a user's direction counts as authorized access. Its outcome will shape whether agents can shop your store freely or only where you let them in, which is a decision merchants will increasingly make on purpose. I dig into the stakes in the legal precedent post.
Care, but do not panic. Agent-driven shopping is real and growing, but it is early and contested. The right move is low-cost insurance: get your structured data and feeds clean, make sure an agent can read your catalog and complete a purchase, and track AI referral traffic. That prepares the surface without betting the company on one browser. The same trade-off shows up across the UCP versus ACP comparison.
Getting your store ready for the agent layer?
I helped build and scale the Shopify Partner Program and scaled consumer brands to nine-figure revenue. I help DTC teams and app founders prepare their catalog, checkout, and analytics for agent-driven shopping without betting the company on a single browser or protocol. If the front-door shift is on your mind, that is the work I do.
Start a conversation See the case studies →A note on sources: Perplexity's June 2026 raise of about $200 million at a valuation near $20 billion to scale Comet is from contemporaneous reporting, including Dealroom and TechTimes. The March 2026 preliminary injunction blocking Comet from parts of Amazon, and Perplexity's April appeal, are from CNBC and PYMNTS. The read on the front door, what changes for merchants, and the four-move checklist are mine, from building the Partner Program and running consumer brands in this ecosystem.