#283 – Shopify Spring ’26 Edition, Faire vs Amazon Business, & Uber customer data

by Paul Drecksler | Jun 22, 2026 | Recent Newsletters

Hi Shopifreaks

This is one of my biggest editions of the year, covering everything from Shopify's Spring '26 Edition to Faire's pivot to serve business-use customers. Let's keep this intro incredibly short today and dive right in…

In this week's edition I cover:

Highlights from Shopify's Spring '26 Edition
Faire welcomes business-use buyers
Uber extends its first-party data off-platform
Anthropic is being sued for not delivering promised tokens
New BNPL rules are coming to the U.S.
Meta launches AI Mode search
Meta expands livestream shopping
OpenAI hints at upcoming ad creative tools
Shopify shareholders voted down a formal AI policy
Amazon launched Brand Elevation
LGBTQ+ consumers are voting with their dollars
TikTok mostly sells things you wear, apply, or consume

All this and more in this week's 283rd Edition of Shopifreaks. Thanks for subscribing and sharing!

Stat of the Week

OpenAI burned through $3.7B in cash in the first quarter of 2026, more than half of its $5.7B in revenue, with both figures roughly tripling from a year earlier, according to documents the company shared with shareholders and reported by The Information. It ended the quarter with more than $73B on hand after its March funding round, a cushion big enough that it may not need to raise again soon or rush to go public. The company posted a $9.3B operating loss, though its gross margin rose to 39% from 33% a year earlier, suggesting the core business is getting more efficient as it scales.

1. Highlights from Shopify's Spring '26 Edition: Selling everything, everywhere, all at once

Shopify released its Spring '26 Edition, dubbed The Everywhere Edition, which is a great name for it given that the edition features “150+ updates that put your products everywhere people are buying.” A recurring theme that I've written about extensively in 2026 is how Shopify is turning the entire Internet into a marketplace for its merchants, and that theme is demonstrated throughout this edition.

If you're unfamiliar with Editions, twice a year, Shopify releases a mega showcase of its latest features called Shopify Editions. I've covered all of their Editions including Winter '26, Summer '25, Winter '25, Summer '24, Winter '24, Summer '23, Winter '23, and Summer '22 (the first one).

Here's what's new in Spring '26 Edition. Some of these updates you've seen me talk about in recent editions and others were announced for the first time:

Universal Commerce Protocol opens to all developers – The centerpiece of the edition is that UCP, the open agentic-commerce standard Shopify co-developed with Google, is now self-serve for any developer. Shopify merchants are UCP-enabled by default, so any surface built on it inherits their checkout rules, discounts, and customizations. All developers have to do is register an agent profile and call a public MCP endpoint to build a full flow from product search to checkout. This is what I meant by Shopify making the entire Internet their marketplace.
Shopify Catalog as the discovery layer – Catalog is Shopify's structured product dataset, spanning billions of products, that AI agents can search and understand. Eligible merchants are in by default, with product details auto-syndicated to ChatGPT, Copilot, Google AI Mode, Gemini, and the Shop app with no third-party apps or manual feeds required. Shopify claims AI searches powered by Catalog convert at 2x the rate of those using scraped data. New this edition: image search, product lookup by URL, richer attributes like size and color, and Shop sign-in for personalized results.
Agentic Storefronts dashboard in the admin – Merchants can now manage all their AI channels from one place in the admin and see orders, sales, and conversions from ChatGPT, Copilot, Google AI Mode, Gemini, and Shop in a single view. The dashboard also shows which AI queries in your category you don't rank for and offers Sidekick fixes when products surface in AI conversations but don't convert.
Catalog API will pay developers – Developers and partners will soon be able to earn revenue when their Catalog-powered experiences drive sales, with sponsored product placements as the mechanism. This is still in developer-preview mode for now, but it'll eventually create a new way for developers to earn revenue from Shopify, outside of app and theme development, which are crowded spaces.
Sidekick App Extensions – Shopify's AI assistant now connects to third-party apps directly, so merchants can ask Sidekick about their tools and take action without leaving the admin. Integrations are live for 15+ partners at launch, including Klaviyo, Loop, Smile, and Judge.me, allowing merchants to get answers from Sidekick in one conversation / dashboard, instead of having to bounce between admins.
Sidekick everywhere – Sidekick is now on every screen in the Shopify app via typing or voice with no full-screen takeover, can run from a phone, and can even pull business insights from an Apple Watch. Sidekick Pulse surfaces proactive next-best-actions on a redesigned admin home built around the assistant.
Campaign Autopilot – A new AI marketing agent that plans, runs, and optimizes paid and organic campaigns across Facebook, Instagram, Shop, and email automatically, within budget and guardrails the merchant sets. This is one of the most standout updates in my opinion, as it puts agencies on notice. More channels coming soon including Microsoft Advertising, ChatGPT Ads, and Snapchat.
Shop Campaigns expands to ChatGPT, Pinterest, and the open web – Shopify's pay-on-conversion performance marketing product, which targets customers at a set price across Shop and other channels like Meta, Google, X, Bing, and Snap, now adds ChatGPT, Pinterest, and open-web advertising via Microsoft Monetize. Merchants only pay when the customer converts.
AI sales associate in Shopify Inbox – A new AI assistant lives on the storefront and answers buyer questions, suggests products, and handles order inquiries using the merchant's own catalog, inventory, and policies. For buyers signed in with Shop, it personalizes recommendations in chat, with merchants able to control its tone and when a human gets looped in.
Shop Pay opens to any brand on any platform – Brands not on Shopify can now offer Shop Pay at checkout, gaining access to Shopify's 250M+ shopper base and one-click purchasing. I better check my PayPal stock price…
Online-to-in-store via the Shop app – Retail locations now surface in the Shop app, product pages show local inventory for in-store pickup, and returns started in Shop finish at the counter with a QR code. This is a great update for omnichannel retailers.
POS v11, Shopify's fastest-ever point of sale – Shopify says its rebuilt core selling flow makes complex carts up to 73 seconds faster, keeps the cart on screen through the entire transaction, and opens discounts, edits, and customer lookups in a side panel. The update is paired with new Verifone Victa mobile hardware that doubles as a countertop terminal.
AI Mode arrives in Shop Search – Buyers can now search Shop conversationally with prompts like, “I need a waterproof jacket for a weekend hiking trip,” and get results tailored to their history, with recommendations sharpening over time. Everyone's using the name “AI Mode” apparently (see Story #6).

There's an ongoing joke (or fear?) in the Shopify world about how “Editions” are “D-Day” for some app developers, as Shopify brings certain features in-house that were previously only available via a third-party app. Now it seems that marketing agencies, advertising SaaS tools, and customer service and AI sales solutions aren't safe either! It's always an exciting, yet scary time when a new Edition comes out…

You can check out the rest of the updates in the full Spring '26 Edition. Which are most impactful for your business? Hit reply and let me know.

2. Faire opens its wholesale platform to business-use buyers

Faire, the wholesale marketplace partially owned by Shopify, opened its wholesale marketplace to business-user buyers, which include businesses like hotels, medical practices, and event planners buying wholesale products for their own operations rather than for resale.

Since launching in 2017, Faire has catered to retailers, who purchase items wholesale to resell in their own stores. However, the company says that there's been “a large, underserved segment of wholesale demand” from these business-use buyers that it's now aiming to cater to.

Faire co-founder and CEO Max Rhodes wrote:

“Businesses come to Faire looking for something that helps them stand out, products that create a feeling and reflect their creative vision. Opening the platform to these buyers is a step toward a future where more businesses, whether stocking their shelves, designing a space, or bringing a menu to life, can easily discover distinct products to build thoughtful, curated experiences for the people they serve.”

The platform aims to serve business-use buyers like:

Boutique hotels stocking rooms with candles, robes, and minibar goods.
Restaurants, cafes, and bars sourcing specialty sauces, pantry staples, glassware, and table linens.
Wellness, fitness, and medical practices stocking skincare, towels, and items for treatment rooms.
Offices and corporate buyers building client gift boxes, employee welcome kits, or stocking snacks and coffee.
Event planners and entertainment spaces purchasing decor, party favors, paper goods, and supplies in bulk for weddings and corporate events.

Faire says that many of its brands are already serving business-use buyers, but have to do so through cumbersome outside workflows, whereas now, they can capture that demand on Faire's platform, which is already integrated with the rest of their tech stack.

Brands are not required to sell to business-use customers, but if they do, the buyers are clearly labeled and must meet the same brand minimums, and brand owners still maintain approval over every order.

I think it's a brilliant move! Why let Amazon and Walmart have all the B2B fun? Chances are, Faire brands are selling on those B2B platforms anyway, and now this gives Faire the ability to take a piece of that market leveraging their existing seller base.

🔥 Partner News

Seguno is officially part of Shopify's Sidekick app extension launch featured in the Spring '26 Edition. The integration lets merchants ask Sidekick questions about their Seguno email data, like how many subscribers haven't opened in the past 30 days, and get answers without leaving the admin. Merchants can also build campaigns through chat, asking Sidekick to draft something like a win-back email featuring a Spring Bestsellers collection, which then opens directly in the Seguno editor to tweak and send. Seguno joins a select group of launch partners that includes Klaviyo, Judge.me, Smile.io, Yotpo, TikTok, Abra Promotions, and AfterShip.

3. Uber extends its first-party data to Meta and Google for real-world ad targeting

Uber Advertising launched a set of “Offsite Ads” that for the first time extend its first-party data signals beyond its own apps to reach consumers on Meta and Google Shopping, while tracking the results back to Uber Eats orders and store visits. This is the first time that Uber has looked beyond its own apps to sell ads, and comes at a time when its ad business is already surging, having surpassed $2B in ad revenue last year.

Here's how it works:

Brands buy Uber ad placements that run on Meta and Google Shopping rather than inside Uber's own apps, extending Uber's reach to where consumers already spend their time online.
Uber targets those ads using its first-party data such as what people actually order, when they order it, and where they were at the time, which are signals it has never made available outside its own ecosystem until now.
Uber then closes the loop by tracking whether someone who saw an ad actually placed an Uber Eats order or visited a store, tying ad spend to real purchases rather than clicks.
One thing to note is that Uber isn't actually handing Meta and Google its data. It keeps the raw signals in-house and only applies them to target the ads, so the platforms never get their hands on your order history. Then again, there's probably a lot that Meta and Google can infer from the targeting signals.

The move turns Uber's ride and delivery data — covering meals, errands, events, and destinations — into a targeting layer brands can use off-platform.

Yippee for customers! Advertisers used to just follow you around based on your web activity, now you'll start seeing ads based on your real-world activities too. Let's hope for your sake the local strip club doesn't advertise on Uber. “Honey, why is Google TV advertising a 50% off buffet coupon for your tenth visit to The Trophy Club?”

Alongside the launch, Uber unveiled new ad formats, including spots that appear within the Uber ride-sharing app, allowing brands to offer sponsored discounts to riders and brand takeovers on Uber Eats. For example, one new in-app format offers a shoppable menu carousel that lets users add items straight to the cart.

The move pushes Uber deeper into the commerce-media race against DoorDash, which rolled out its own full-stack advertising expansion just a few days earlier.

4. Anthropic is being sued for delivering fewer tokens than it advertised

Using Anthropic and OpenAI is like writing a blank check and depositing it into a black box. You're authorizing spend you can't see, against limits you can't track, on work you can't audit for efficiency. That's the basis of a proposed class-action lawsuit by a Claude user who is suing Anthropic over its $100 Max 5x and $200 Max 20x coding tiers allegedly delivering far below the advertised capacity.

The plaintiff, Karl Kahn, says he blew through his caps almost immediately after upgrading, once burning 15% of his entire weekly allowance in a single five-hour session.

Here's why the lawsuit has merit:

There's no live meter. When using the Claude app, nothing tells you “this response cost X tokens” or “you're at Y% of your weekly cap”. You find out you're at the end of the token road when you hit the brick wall.
You can't get estimates. At no point does Claude give you any indication of how many tokens you're about to use before it begins burning them for a task.
The caps are stacked, not simple. Usage runs against a five-hour rolling window AND a separate weekly cap at the same time. Anthropic employs two overlapping meters to track your usage, neither of which you can see.
You can't tell efficient work from wasted work. If a task takes three tries instead of one, Claude burns more tokens, but you can't tell whether that's because the task was genuinely hard or because the model did something wasteful.
Claude fucks up all the time, and you still have to pay for it. Your tokens are burned whether or not Claude is successful at a particular task, and oftentimes it's unsuccessful behind the scenes before producing a final output, which means you never even know about the wasted token usage.

The lawsuit targets whether Anthropic delivered upon its advertised usage limits. However, the real question is less about whether it's delivering on all promised tokens, and more about whether it's efficiently using those tokens. Anthropic is already moving away from unlimited and capped plans toward usage-based plans, so the “5x” or “20x” become irrelevant, but the question over how Claude uses your tokens remains.

When AI companies charge based on usage, how efficient they are with your tasks becomes a YOU problem, not a THEM problem. Even worse, there's a structural incentive for usage-based models to be inefficient, as it drives up revenue.

I'm not claiming that AI companies are intentionally inefficient to drive up revenue, at least right now in 2026, as most are still subsidizing usage. However, at some point in the future, the question will change from “How efficient CAN we be?” to “How efficient MUST we be?” — especially when single-digit efficiency margins of error on token-usage, which may be indistinguishable to the user, equal billions of dollars a year to the company.

What are your thoughts? Hit reply to this e-mail or join the conversation on my LinkedIn post.

5. New BNPL rules are coming to the U.S. one way or another

New York Democrat Representative Dan Goldman introduced a bill that would give BNPL users the same legal protections credit-card holders get, an area where they currently have no federal safeguards. The Buy Now, Pay Later Consumer Protection Act would require BNPL providers to spell out fees clearly, send periodic statements, give borrowers a window to pay before they're hit with late fees or dinged on their credit, and grant dispute rights like refunds on returns and protection from unauthorized charges.

The bill follows the Trump administration's reversal of a 2024 CFPB rule that would have regulated BNPL loans like credit cards, which had left users without federal protection. Goldman said the bill responds to more Americans leaning on BNPL for everyday costs, taking a stab at President Trump by saying, “In Trump’s nosediving economy, Americans are increasingly being forced to use Buy Now, Pay Later to cover basic expenses.”

If this sounds familiar, it's because the bill is very similar to one proposed in December 2025 by U.S. Senators. That one was called The Buy Now, Pay Later Protection Act, while this new one from Goldman is called The Buy Now, Pay Later Consumer Protection Act. LOL, the names keep getting longer, but the lack of consumer protection in the space remains the same.

Meanwhile, states aren't sitting around waiting for Congress to figure their shit out (because that might be a while)…

This past March, New York unveiled the nation's first comprehensive regulatory framework for BNPL lenders, stepping in on a state level to fill the regulatory gap that the CFPB left.
Last week, Illinois passed its own bill requiring BNPL lenders to register with the state, disclose terms, and weigh a borrower's ability to repay before extending Pay in 4 and other point-of-sale installment loans. The bill now heads to the Governor for approval.
In 2023, California determined pay-in-four products were loans under its lending law and required providers like Afterpay and Klarna to get California Financing Law licenses, which was a step in the right direction, but not quite a comprehensive consumer-protection package like the one New York built.
Oregon and Washington have also actively pushed BNPL operators to obtain licenses, treating the pay-in-four model as consumer lending regardless of whether interest is charged.
Massachusetts uses its existing small loan and criminal usury statutes to monitor and penalize out-of-state fintechs that don't comply with state interest and fee caps.

While Congress may be dragging its feet with nearly identical bills across the Senate and House of Representatives that inevitably get nowhere because they're all still on vacation, states are starting to step up with their own regulation. I'll keep you posted as more dominos fall.

6. Meta gets serious about search (again) with AI Mode

Meta introduced AI Mode, a new search tab powered by its Muse Spark model that answers questions using information curated from public posts within its apps instead of the broader web. Muse Spark is the company's first major model developed by Meta Superintelligence Labs, which is headed by former Scale AI CEO Alexandr Wang, and debuted in April.

The company wrote:

“AI Mode is a new way to get answers to your questions right on Facebook thanks to Meta AI. From exploring your Feed to searching for something specific, AI Mode uses Meta AI to give you answers grounded in what people are saying publicly across our apps like in Groups and Reels, so you get real perspectives and experiences rather than a generic list of search results.”

Yes, it's also using the same “AI Mode” name as Google and now Shopify. You're not imagining things.

I think it's a great idea for Meta to populate answers from its own trove of user generated content. It's the biggest way Meta can stand out in the search world, and it's no different from what Google is doing in partnership with Reddit. However, some are questioning the integrity of its answers and whether we should trust information generated from your crazy uncle's Facebook posts.

TechCrunch wrote:

“Both AI Mode and Forum’s Ask tab raise a familiar question: How reliable are answers generated from public posts and group chatter? Because the AI is summarizing content from everyday users rather than vetted sources, there’s a real risk of outdated or misleading information slipping through.”

The Verge wrote:

“Things got frustrating with more specific requests. I asked for things to do nearby, and it suggested a swim at the community pool — noting it would be closed over the weekend. It cited a post on the pool’s Facebook page, and the hours listed on the same page. But when I checked the source, nothing suggested that the pool would be closed over the weekend — and the cited post doesn’t seem to exist. The pool’s website also confirms that it’s open Saturday.”

Though to be fair, are Gemini, Claude, or ChatGPT's sources any better? Many of the answers that surface for me on those AI platforms are pulled from Reddit or other community forums anyway. At least Meta has information about the profile that provided the post, a reliability data point that those other platforms don't have.

Meta's AI Mode won't be perfect — no AI is. However, it's about time that Meta got serious about search, again, especially if they plan on pushing deeper into commerce, again. On that note…

7. Meta expands livestream shopping with ads, virtual cards, and affiliate deals

Meta is leaning harder into livestream shopping, putting its live-video shopping ads on Instagram and rolling them out worldwide on Facebook. New Facebook tools let viewers shop promoted items mid-broadcast without leaving the stream, powered by live-shopping platforms like CommentSold, Firework, and TalkShopLive.

Meta also introduced virtual cards for shoppers, which are temporary numbers tied to the Mastercard or Visa they already have, so they never hand a merchant their real card details. (Meta's like, “We know we have scams on our platform, so here's a virtual card so that you only get ripped off by this one purchase instead of having your credit limit maxed out.”)

On the creator side, Meta is expanding its affiliate partner program, allowing creators to earn commissions by tagging products from India's Flipkart, Latin America's Mercado Libre, and soon Lazada across Asia.

Lastly, the company is automating product ad creation. Advertisers can simply supply raw product data and creative assets, and Meta's AI system builds a tailored ad format for each individual viewer on the fly.

Remember in 2022 when Facebook quit livestream shopping and Instagram shut down its affiliate program? Lock in for a second because I'm about to dump hard on Meta…

It's amazing to me that Meta is as successful as it is because it's got to be the most poorly run Big Tech company I've ever seen, with absolutely no long term vision. Meta's entire empire has been built by acquiring, copying, or stealing. The company isn't intrinsically cool or culturally relevant — it simply buys both. Nothing demonstrates this more than Meta's on-again off-again relationship with commerce.

2015 – Facebook began testing “Buy” buttons and early Shop sections on Pages, its first real push into on-platform commerce.
2016 – Facebook Marketplace launched, which is about the only thing they ever got right with commerce. Of course, they took the idea from Craigslist.
2017 – Instagram introduced shoppable post tags, letting brands tag products in posts.
2018 – Facebook debuted Live Shopping, its first livestream-commerce feature.
2019 – Instagram launched Checkout, enabling native in-app purchases without leaving the app.
2020 – Meta launched Facebook Shops and Instagram Shops, built via partnerships with Shopify and BigCommerce. Later that year, they aggressively forced adoption by replacing Instagram's notification tab with a home-screen Shop Tab, and expanded Live Shopping to all U.S. sellers.
2021 – Meta partnered with Shopify so merchants could sync catalogs into Shops, and Instagram began testing affiliate commerce. Later that year, they added Live Shopping Events and “Drops” for product launches.
2022 – Instagram shut down its affiliate-commerce program. Facebook shut down Live Shopping.
2023 – Instagram ended Live Shopping entirely to wrap up Zuckerberg's “year of efficiency.” Meta dropped the Facebook Shop tab, sunset standalone Shops in most non-U.S. global markets, and formed a surprise alliance with Amazon allowing users to link accounts and buy directly from Amazon feed ads.
2024 – Meta implemented a strict “Native Checkout or Die” mandate for U.S. merchants, forcing them to process payments on-platform or lose their shop and product tagging features. Meanwhile, they leaned hard into AI-driven shopping ads like Advantage+ while TikTok Shop surged.
2025 – The native checkout strategy completely failed due to merchant backlash and operational complexity of running checkout. (Sound familiar? *cough* OpenAI…) In a massive U-turn, Meta deprecated native in-app checkout and forced everyone back to an external-link model, transforming Shops back into visual discovery galleries that route the transaction back to the merchant's own website.
2026 – Meta re-enters commerce discovery via “AI Mode” search, scraping deep product catalog feeds to answer conversational buyer queries directly in chat. They also resurrect livestream shopping by partnering with third-party software (like CommentSold and TalkShopLive) to power “Live Video Shopping Ads” and comment-to-buy DMs, leaving the transaction logistics to external providers.

It's no wonder none of Meta's commerce efforts ever took off! They didn't give brands, merchants, and consumers time to adapt.

Imagine where Meta would be with e-commerce, livestream shopping, and affiliate commerce if they had stuck with it for the past decade. If they had created the future of social commerce, instead of copying it.

8. OpenAI terms of service hint at upcoming ad creative tools

OpenAI is moving to automate ad creative, the most labor-intensive part of advertising, with updated terms revealing AI tools that let advertisers generate, modify, optimize, and localize their assets, according to Digiday.

For the pilot's first four months, OpenAI required advertisers to upload their own creative, whereas the new tools hint at handling more of the production. Advertiser beware though, because its policy disclaims responsibility for errors in what the AI generates and leaves the task of reviewing the output to the advertiser (which is fair in my opinion).

The company's updated policy reads:

“OpenAI may make available AI-powered Creative Tools that allow you to generate, modify, transform, optimize, localize, or translate advertising creatives using Ad Materials… OpenAI is not responsible for errors, omissions, outdated information, or inconsistencies in Ad Materials or for Claims or losses arising from Generated Creatives that you approve or use”

The tools would put OpenAI on par with Meta and Google, which offer AI creative generation through Advantage+ and Performance Max, and in direct competition with platforms that specialize in AI ad creative generation including AdCreative.ai, Pencil, and Cuttable.

I just hope that their ad generation tools are better than what we've previously seen from OpenAI! In February, after ChatGPT ads started to appear for more users, I roasted the company on its poor execution with duplicate ad verbiage, cut-off text, and generic sounding ads.

9. Other e-commerce news of interest

Shopify shareholders voted down a proposal calling on the company to adopt a formal AI policy, with just under 14% backing it at the company's annual general meeting. The Shareholder Association for Research and Education (SHARE) brought the measure on behalf of the United Church of Canada's pension plan, arguing that generative and agentic AI carry risks around human rights, misinformation, and erroneous automated transactions. Shopify's board urged a no vote, calling the proposal disconnected from how it builds technology and saying its existing contracts and terms of service already set guardrails around AI, though proponents of the AI policy say those contracts don't provide enough transparency into Shopify's AI risk management and governance controls. SHARE said it expected to lose the vote, given Shopify's dual-class share structure and the founder share that gives CEO Tobi Lütke 40% of the voting power on top of his Class B stake, which makes any shareholder proposal the board opposes nearly impossible to pass. The news follows my report from earlier this month about how Shopify wants governments to stay out of its way with AI regulation.

Amazon launched a new program called Brand Elevation that gives brand-submitted data priority over third-party sellers in how products are described and stops them from creating new catalog listings for items unless the brand authorizes them to do so. Amazon says that the new offering gives brands a “competitive advantage on the detail page” and “eliminates catalog errors” caused by incorrect data from third-party merchants or its systems overriding brand-submitted information. However, The Information framed the move as Amazon giving brands “priority over independent merchants” as it courts larger brands like Nike and Estée Lauder to its platform. I think that the change is long overdue and will benefit consumers and brands alike by creating a single source of truth for product information and helping customers avoid counterfeit merchandise. That said, I do understand the limitations it puts on resellers, who effectively now need a brand's approval to sell items that the brand itself doesn't list on Amazon, which may push them to other marketplaces.

Google launched an AI tool called “Ask Ad Manager” that helps publishers troubleshoot campaigns and cut down on slow, repetitive manual work involved with selling and managing ad space on their sites. Initially the tool covers three problem areas: 1) diagnosing why a campaign is underperforming, 2) answering detailed questions about how specific bidders are doing and how campaigns compare to industry benchmarks, and 3) steering publishers to suggested fixes or the data they need. Trials started in mid-June with publishers including Yahoo, which built Ad Manager into its own agents, with Google planning to ship developer tools later in the year to support more workflows. For now, Ask Ad Manager doesn't take any automated action and a human must apply any suggestion, but Google frames it as an early step toward a broader agentic advertising system where AI handles inventory, pricing, and negotiations.

LGBTQ+ consumers are steering their spending away from Target, Walmart, and Amazon over what they see as those companies' retreat from diversity commitments, according to a new survey from the Human Rights Campaign Foundation. Nearly 72% of LGBTQ+ respondents said they buy fewer products from brands they view as pulling back on DEI, with Target, Walmart, Amazon, Chick-fil-A, and Home Depot named most often. Conversely, 70% said they bought more from brands they consider inclusive such as Costco, Apple, Ben & Jerry's, Delta, and Kroger. LGBTQ+ consumers represent more than $1.4T in annual U.S. spending power, according to the HRC Foundation's estimate, which is around 7% of annual consumer spending in the U.S. One important thing to note though is that the survey respondents indicated that they bought “fewer” or “more” from certain businesses, not that they “didn't shop” or “exclusively shopped” at those places, so the true economic impact is only a fraction of that $1.4T buying power mentioned.

Carvana is extending its all-online sales model to new cars, rebranding its Stellantis franchise dealerships as test-drive centers and “playgrounds” rather than places that sell new vehicles. For example, the company's Dallas test location has no finance offices or commission salespeople, just hourly associates and QR codes that let shoppers customize and test-drive vehicles before buying through Carvana's website, where the cars are actually sold. Carvana has spent more than $171M acquiring several Stellantis dealerships and won approval as the automaker's certified website provider, giving it a unique advantage over rival dealers. Its online-only approach breaks from a franchised-dealer model that NADA says topped $1.3T in U.S. sales last year, while Carvana now ranks as the most valuable U.S. auto retailer, with a market cap above $70B — though Amazon is coming for its lunch with Amazon Autos.

Pinterest is expanding its AI-powered Performance+ ad suite and Shopify integration with two updates aimed at SMBs, including a new customer acquisition feature and a one-click campaign setup for Shopify merchants. The new customer acquisition update lets advertisers feed in their existing customers so its AI can steer bids and delivery toward similar high-intent shoppers, with the option to weight new buyers more heavily through value rules. Pinterest says early testing of the tool, which reminds me of look-alike audiences on other platforms, raised new-customer conversions by 64% on average, which is a pretty big lift! The second update lets eligible merchants launch a Performance+ shopping campaign directly from Pinterest's Shopify integration, with best practices enabled by default, and is part of Pinterest's broader push to pull more small businesses onto the platform.

TikTok Shop's top sellers are overwhelmingly products people wear, apply, or consume, driving over two-thirds of revenue across the 1,000 best-selling U.S. products, according to Marketplace Pulse. Three categories take most of the sales with beauty and personal care at about 30%, health and wellness around 20%, and apparel near 17%, leaving home goods, electronics, automotive, pets, books and the rest to split what's left. The median top product sells for about $40 and roughly two-thirds of items are priced under $50, with 1% of sellers driving about 60% of GMV on the platform. Marketplace Pulse notes that 13 of the 20 biggest sellers from TikTok Shop's first full year still rank among the 1,000 top sellers, including medicube, Halara, MERACH, Goli, tarte, and MaryRuth's, which together make up about a quarter of tracked branded revenue.

Amazon is testing software that reassigns warehouse workers to different sections of the facility in real time as package volumes shift in a similar way it's automated the routing of packages throughout facilities, according to documents seen by Business Insider. The tool, called Full Facility Load Balancing, rechecks staffing every three minutes or so and flags workers to move from overstaffed spots to busier ones, which the documents project could save nearly 7M labor hours and $193M a year — while helping to ensure that employees never get a single moment of reprieve while on the clock. Amazon disputes those numbers, calling them inaccurate and drawn from hypothetical models, not real productivity, and says managers stay in charge while the tool just feeds them faster information. The pilot targets Container Build, which is where workers load packages into outbound carts, and aims to reach every robotics-enabled fulfillment center it runs in North America this year.

LLMs.txt, the file many publishers have adopted to help get their content surfaced by AI, was never built for that purpose and won't deliver it, according to Google's John Mueller. On Google's Search Off the Record podcast, Mueller said that one of the creators of LLMs.txt told him it was originally meant to help an AI that already knows a site find more of its pages, not to make a site discoverable. Okay, so it helps an AI discover more of a site's pages? I feel like we're debating semantics here. He also called the format inherently untrustworthy, since it lets a site owner simply assert what their content is about, something an LLM has no reason to believe, and stressed that discovery and ranking still run through a site's HTML. Mueller pointed to Google's WebMCP as being a better fit for e-commerce, since it gives AI agents specific abilities like filtering products, comparing them, and adding items to a cart, though he said that no agentic standard has won out yet.

TikTok Shop sellers have been luring shoppers into livestream auctions with iPhones and iPads, then awarding most winners cheap prizes like teddy bears and charging cords, according to a WIRED investigation. Through a feature called Surprise Sets, which was added to TikTok last year, auction hosts create buckets of up to 500 prizes, and whichever viewer bids the most walks away with a random prize from the lot. The problem is that aside from a handful of high value items, most of the prizes are junk, which has pissed off a lot of buyers. WIRED reached out to TikTok for comment during the investigation, and one day later, the company rolled out a change to its livestream policies that prohibit hosts from including iPhones, iPads, televisions, diamonds, gift cards, or precious metals as part of the prizes available during a Surprise Set, a change that TikTok claims was already in the works. TikTok bans gambling and says it does not consider the auctions gambling, but what the fuck else are you supposed to call it when you pay money to spin a wheel and win a prize?

Speaking of gambling… Polymarket paid creators to stage fake winning bets on lookalike versions of its website, according to a Wall Street Journal investigation, which found that none of the roughly $1.9M in bets shown across 1,105 videos were genuine. The campaign cuts against the company's core pitch that every real trade settles on the public Polygon blockchain in USDC where anyone can audit it, since the staged trades ran on dummy sites that no ledger could verify. One clip showed a creator winning $100,000 on a bet that Trump said “McDonald's” in January, an event that never happened, while public data shows more than 50 real accounts placed that bet and all lost. Creators, who were mostly college-aged, earned $2,000 to $3,000 a month and were told not to disclose the payments, while Polymarket hired a marketing firm to push the clips past 140M views. So an off-shore gambling website that's riddled with insider trading faked bet outcomes to grow in popularity? Color me surprised…

BNPL is coming to video games and ride sharing! Wait, why did I use an exclamation point? That's not really a good thing… Xbox appears to be preparing a BNPL option for its products, according to code found by a data miner on Xbox's website that names PayPal and Klarna. Microsoft neither confirmed nor denied the news, nor did they provide any clarification about whether BNPL was potentially coming to video consoles, video games, or both. Meanwhile in Europe, Klarna has partnered with Bolt, a European ride-hailing app, to integrate its BNPL payment options across Sweden, Germany, Finland, and Norway. Personally, I don't think video games or ride-sharing should be so cost-prohibitively expensive that customers need to pay-in-4, but now they can. Though to be fair, Klarna positions the integration as part of its push into Pay in Full methods.

Did you know that someone made a documentary about the scandal involving eBay executives cyberstalking Ina and David Steiner of EcommerceBytes? I had no idea! Liz Morton of Value Added Resource, who has been covering the scandal for years, wrote that “Whatever It Takes: Inside the eBay Scandal,” which is streaming on Amazon Prime of all places, omits the context that would make it more than a “few bad apples” story. Morton said the film never names SVP Wendy Jones, the security chief's direct supervisor and later a defendant, nor the anonymous critic “unsuckEBAY” whom executives sought to unmask, and framed the case as a “workplace vendetta” rather than a corporate attack on journalists and protected speech. She also pressed on governance threads the film sidesteps, including the 8-figure exit packages for departed executives that eBay declined to claw back, founder Pierre Omidyar's silence, and the film's own ties to Omidyar-adjacent media networks. Whether or not the documentary falls short, it sounds scandalous! I'm going to need to give it a watch.

Amazon's Prime Air drone delivery sparked protests in Richardson, Texas, where residents say the 80-pound drones crossing over their homes and schools have turned the sky into a “drone highway,” or as Kenny Loggins likes to sing, “a highway to the danger drones!” Neighbors are complaining about the constant noise from low-flying flights they estimate at barely 100 feet that pass by every few minutes, with one person saying she can hear them in her home even with the TV on. Residents also say they got little warning before the program launched and that complaints they've brought to the city and the FAA have brought no results, while Amazon contends that customer reaction has been strongly positive and that it's responded to concerns, such as raising altitudes and rerouting some flight paths.

Kroger's e-commerce operations were profitable for the first time in its most recent quarter, with digital sales up 19% YoY, after grocery delivery drew a “record number” of new customers to its e-commerce platform. Notably, orders delivered in less than one hour represented about half of Kroger's e-commerce growth during the quarter. Outside of e-commerce, the company's retail media business grew more than 20%, while comparable sales excluding fuel rose just 1%, down from 3.2% a year earlier, total sales rose about 2% to $46.1B, and operating profit grew more than 6%. Kroger's new CEO Greg Foran, who previously led Walmart's U.S. business, said about 60% of Kroger's stores need to improve and that the chain has grown its footprint too slowly against rivals, but expects e-commerce to continue expanding.

In lawsuits this week…

Amazon is facing a possible FTC lawsuit over claims it misled advertisers about the terms and pricing behind the sponsored listings that sit above (and between, and below) its product search results, according to Bloomberg sources. The case could carry billions of dollars in civil penalties due to the involvement of several state attorneys general, whose consumer protection laws allow daily fines beyond the federal limits on what the FTC can collect on its own.
TikTok is being sued by Florida Attorney General James Uthmeier for allegedly violating HB3, the state's online child-protections law, by letting children under 14 create accounts and admitting 14 and 15 year olds without parental consent. The complaint also accuses the app under Florida's deceptive-practices act of using design features built to hook kids and misleading parents about its sexualized, drug-related, and mature content. Uthmeier is seeking financial penalties and an injunction forcing compliance, while TikTok says it's engaged in good faith and has already begun suspending under-14 accounts in the state.
Meta may face a revived antitrust suit after a group of Facebook users petitioned the 9th Circuit to reinstate their dismissed case alleging that the company monopolized the market for “personal social network services” and used that dominance to harvest user data without paying for it. The plaintiffs argue that a jury rather than a trial judge should have weighed their estimate that, in a competitive market, Meta would have to pay users about $5 a month for their data. The suit runs parallel to a separate FTC antitrust case Meta also defeated and is now appealing.
xAI lost its trade-secret suit against OpenAI after a federal judge dismissed it with prejudice, ruling Elon Musk's company failed to show that OpenAI improperly obtained confidential information about its Grok chatbot and that asking a former xAI engineer about his prior work during recruiting was routine hiring, not inducement to reveal secrets. The dismissal marks Musk's second courtroom loss to OpenAI in four weeks, after a jury in May rejected his $150B suit alleging OpenAI abandoned its nonprofit mission.
Meta failed to get an adult-film copyright suit dismissed, with a federal judge ruling the producers don't have to prove their films trained Meta's AI because the downloading itself is the infringement. The judge also rejected Meta's claim that employees grabbed the content for personal use, pointing to 47 corporate IP addresses pulling thousands of files between 2018 and 2025 as evidence of systematic collection.
Meta must face a copyright suit from Eminem's publisher Eight Mile Style over allegedly hosting 243 of the rapper's songs without a license across Facebook, Instagram, and WhatsApp. The publisher is seeking $150,000 per work x 243 songs x 3 platforms, putting the total potential damages at around $109M. It'd be hilarious if Meta argued that “Facebook, Instagram, and WhatsApp” are part of just “one platform called Meta” to save money, and then later that argument came to bite them in the ass in a different lawsuit!
Speaking of Meta copyright infringing songs… Meta asked a federal court to throw out another copyright and defamation suit, this time from Wixen Music Publishing, which is seeking over $102M across 681 songs. Meta is attempting to frame the case as a routine licensing dispute, after a deal between the two companies lapsed in December 2025, but Wixen counters that Meta cut rates to swap songwriter-made music for royalty-free AI-generated tracks in an effort to squeeze independent publishers holding out for fair pay. Without any information whatsoever about this case, I'm going to side with Wixen.
Last but not least in music publishing lawsuits… X asked a federal court to throw out the NMPA-led music publishers' copyright suit accusing it of infringing more than 2,000 songs and seeking up to $150K each, a case which dates back to June 2023 when the platform was still Twitter. (To many, it is still Twitter. LOL) X argues that the Supreme Court's March Cox v. Sony ruling shields it from infringement, an argument Meta is using in some of its cases too.
Three Amazon software engineers filed a civil-rights complaint alleging the company retaliated against them for backing data-center regulation before the Seattle City Council, claiming each was pulled into an HR investigation with firing on the table a day after the council passed a 12-month ban on big new data centers. Amazon disputes the allegations, saying it's only checking a possible breach of its policy against speaking for the company without approval and that it never threatened to fire them.

In corporate shakeups this week…

Fiserv named Takis Georgakopoulos CEO effective immediately after Mike Lyons abruptly left to run Truist Financial, making him the company's second chief executive in two years as it fights to recover from an earnings slump that's cut its stock more than 70%.
Noam Shazeer, the Google DeepMind researcher and key author of the original transformer paper behind modern GPT models, is leaving Google's Gemini team to join OpenAI, two years after Google brought him back through a $2.7B deal for his startup Character.AI.
Barret Zoph, OpenAI's head of enterprise AI sales, departed just five months after returning to the company, having rejoined in January following his abrupt exit from Mira Murati's Thinking Machines Lab over alleged misconduct involving an undisclosed relationship with a colleague.
BigBasket, the Tata-backed Indian quick-commerce and online grocery platform, replaced co-founder and CEO Hari Menon with longtime Amazon India executive Amit Nanda as it fends off fierce quick-commerce competition from Blinkit, Zepto, and Swiggy Instamart.
X is trying to poach Meta engineers with a cheeky pitch to “match or even exceed any snack budget offer,” after Meta CTO Andrew Bosworth said the company would upgrade office-kitchen snacks to lift morale following repeated layoffs. Really? Snacks? Meta is so freaking out of touch, it's insane.

Amazon has withdrawn from distributing “Artificial,” a scripted feature film about OpenAI CEO Sam Altman, amid Amazon deepening its business relationship with OpenAI and as the AI firm prepares to go public, according to Puck. The film features Altman and Elon Musk, painting neither in a flattering light, which some people believe may have influenced Amazon's decision. Amazon has not linked the exit to its OpenAI partnership, but the timing is anything but coincidental, as the news just broke earlier this month that OpenAI confidentially filed for an IPO. Well, I hope someone else distributes the movie instead because I want to watch it!

President Trump said he viewed Anthropic as a threat to national security as of last week, but relations have since improved, in an exclusive interview with “The Axios Show.” When asked whether Anthropic or its CEO, Dario Amodei, posed a threat, Trump said “not now, but a week ago, maybe,” and called Amodei nice and smart after the G7 event. His remarks follow a crackdown in which the Commerce Department blocked everyone outside the U.S. from Anthropic's top models and the Pentagon flagged it as a supply-chain risk, set off by an Amazon report on a vulnerability. Anthropic said it's grateful for the administration's partnership in resolving the matter. See my coverage from last week for more details about what led up to Trump's comments.

Amazon is in talks to sell its Trainium AI chips for use in other data centers, according to Amazon VP Peter DeSantis. Until now, the chips have been available exclusively through Amazon Web Services, but now DeSantis says, “We view AI infrastructure as rapidly evolving, and we're constantly looking at ways to get to more customers.” DeSantis dismissed the worry that selling Trainium outside AWS could undercut Amazon's cloud business, citing heavy underconsumption in AI, with the chip's third generation already sold out and a fourth drawing interest. The move follows Google opening its own TPUs to outside data centers in April and builds on CEO Andy Jassy's projection that external sales could roughly double Amazon's annualized chip revenue toward $50B.

Etsy launched a “Shop Other Jeffs” marketing campaign that spotlights the more than 5,000 independent sellers named Jeff on its platform, taking a shot at Jeff Bezos right before Amazon's Prime Day, which starts tomorrow. The campaign features the tagline “one Jeff should not rule commerce” and showcases makers like a North Carolina potter, a Tahoe woodworker, and an Ann Arbor lighting designer, framing Etsy's nearly 6M sellers as a values-driven alternative to convenience-first shopping. The ads roll out across broadcast TV, streaming, paid social, and out-of-home activations in NYC, Seattle, and DC, including, with the irony fully intended, placements on Amazon Prime Video.

🏆 This week's most ridiculous story (and speaking of Jeffs)… Jeff Bezos predicted that AI will create more jobs and even cause a labor shortage rather than make workers redundant, pushing back against fears that the technology will wipe out large numbers of jobs. And when exactly will that happen, Jeffrey? In our lifetimes or in like three generations from now? He argued at a Paris tech conference that people are held back by barriers technology can remove, not by limited ambition, and that AI will unlock new opportunities and increase demand for human labor. The claim contradicts what everyone else is actually seeing happen in the market right now.

10. Seed rounds, IPOs, & acquisitions

Salesforce signed a definitive agreement to acquire Fin, an AI customer-service company formerly known as Intercom whose AI agent resolves support queries across channels like chat, e-mail, WhatsApp, and Slack, for approximately $3.6B. Fin's core product is powered by a proprietary support-tuned model called Apex that the company claims outperforms frontier models from OpenAI and Anthropic on resolution rates, closing roughly 76% of support requests without a human. The deal brings more than 30,000 business customers to Salesforce as it folds Fin into its Agentforce platform.

American Express announced plans to acquire TheFork, an online restaurant reservation and management platform that operates in Europe, from Tripadvisor, for $700M in cash. The deal builds on Amex's earlier acquisitions of reservations apps Resy and Tock, and the company expects the three platforms to collectively cover 75,000 bookable venues. For Tripadvisor, the sale follows months of pressure from activist investor Starboard Value to offload TheFork, freeing the company to refocus on its higher-growth experiences business, Viator.

Yum Brands is selling Pizza Hut for $2.7B, splitting the 68-year-old chain between private equity firm LongRange Capital, which is buying the business outside mainland China for about $1.5B, and Yum China, which is taking the China operations for roughly $1.2B. Pizza Hut has long been the weak spot in Yum's portfolio, with U.S. sales down 8.2% last year, and Yum had already announced plans to close 250 U.S. locations in February. The restaurant chain has faced intense competition from Domino's and other national pizza chains for decades, and more recently from DoorDash, Uber Eats, and other food delivery apps that offered more types of food for delivery than just pizza. Yum said the sale lets it focus on its stronger brands, KFC and Taco Bell.

Respond.io, a customer conversation management platform that unifies WhatsApp, Instagram, TikTok, voice, and e-mail into one inbox, raised $62.5M in a Series B round led by Camber Partners, bringing its total amount raised to more than $69M. The platform launched as a unified messaging inbox and now offers AI agents that qualify leads and handle interactions before handing off to humans, with Meta and TikTok tapping it for early rollouts like WhatsApp's Business Calling API and TikTok Business Messaging. The company will use the funds to expand into North America and Western Europe, after years of building its base across Asia-Pacific, Latin America, and EMEA.

Akeneo, a platform that helps brands centralize, enrich, and syndicate product catalog data across sales channels, acquired PricingHUB, an AI pricing-optimization platform, for an undisclosed amount. The deal lets retailers manage product data and pricing decisions, including promotions, markdowns, competitive intelligence, and AI-driven demand modeling, in one environment rather than across separate systems. PricingHUB will operate as a dedicated business unit within Akeneo, maintaining continuity for its existing customers while the two platforms are integrated.

Odyssey, an AI startup building “world models” that simulate the physical world with realistic physics for uses from video games to robotics, raised $310M in a Series B round led by Natural Capital with participation from Amazon at a $1.45B valuation, bringing its total amount raised to $337M. The company was founded in 2023 by self-driving veterans Oliver Cameron, who sold his autonomous-vehicle startup Voyage to GM's Cruise, and Jeff Hawke, an engineer from UK self-driving firm Wayve. As part of the deal, Odyssey made AWS its preferred cloud provider and will optimize its models to run on its Trainium chips, a competitor to Nvidia's AI chips.

BlueConic, a customer data platform that turns brands' first-party data into real-time customer profiles, acquired Blueshift, an AI platform that runs marketing decisions and delivery across email, push, in-app, and SMS, for an undisclosed amount. The deal pairs BlueConic's onsite data capture and decisioning with Blueshift's cross-channel execution into one stack, which the companies are pitching for agentic marketing, where AI agents act on customer behavior in real time instead of working from data copied out of a warehouse hours earlier. The combined company serves more than 600 B2C brands across CPG, retail, DTC, and travel and hospitality.

TruthEngine, a UK-based online review verification company that uses AI and forensic analysis to detect fake and manipulated reviews, raised an undisclosed seven-figure Series A round led by automotive and technology entrepreneur Tom Wood. The raise comes as the UK's DMCC Act makes review authenticity a legal obligation that sits with businesses, not just review platforms, with the CMA able to fine companies up to 10% of global turnover for breaches. TruthEngine plans to use the funds to expand across sectors like financial services, telecoms, and consumer marketplaces.

Mimir, an Oslo-based startup building an AI platform that automates customer support and broader operations for e-commerce brands, raised $600K in a pre-seed funding round led by Sondo Capital. The company, founded in 2024, handles around 250,000 conversations a month for roughly 60 brands across five countries, with agents that resolve issues end-to-end rather than just creating tickets for human intervention. Mimir is profitable and has grown 7x over the past year without outside funding. It plans to use the funds to accelerate product development and double its team.

Go, Japan's largest taxi-hailing app, raised ¥88.6B ($552.6M) in the country's biggest IPO of the year so far, with the stock jumping as much as 23% above its ¥2,400 offer price before easing to an 8.8% gain after the midday break. Founded in 1977 as a taxi operator and backed by Goldman Sachs-affiliated funds and NTT Docomo, Go competes with Uber in Japan's domestic market and runs its own in-app payment service, GO Pay. It plans to put the proceeds toward developing its robotaxi business and funding expansion, including acquisitions inside and beyond ride-hailing.

L'Oréal acquired a majority stake in Innovist, an India-based beauty company built around online-first brands like Bare Anatomy, Chemist at Play, and SunScoop, for an undisclosed amount reportedly between $350M and $450M. The deal pushes L'Oréal deeper into one of the world's fastest-growing beauty markets and marks its first purchase of an Indian company in nearly 13 years, since it bought Cheryl's Cosmeceuticals in September 2013. Innovist's three founders will stay on as minority owners and keep operating the business with L'Oréal India, with the company's brands moving into L'Oréal's Consumer Products Division and L'Oréal holding the right to buy the founders out fully down the line.

Gradial, an AI marketing-operations platform that builds agents to run campaign work for large enterprises, raised $65M in a Series C round led by Insight Partners at a $675M valuation, bringing its total amount raised to more than $110M in 16 months. The company was founded in 2023 by four Dartmouth classmates, including former SpaceX engineers, and deploys agents across tools like Adobe, Salesforce, and ServiceNow to handle authoring, quality assurance, brand compliance, and publishing, with customers including AWS, T-Mobile, and Vanguard. Gradial's annual recurring revenue has grown more than 10x over the past year, with early adoption concentrated in regulated sectors like healthcare and financial services, where compliance rules get encoded directly into its agents' workflows.

RealTime Reservation, a platform that lets hotels and resorts sell and manage on-property amenities and activities, acquired STAY, a branded guest-experience app that lets hotels offer mobile ordering, service bookings, and digital concierge from guests' smartphones, for an undisclosed amount. The deal expands RealTime's international footprint, with the combined platform serving more than 2,000 hospitality properties across 75+ countries, and gives operators a single platform spanning booking, service delivery, and non-room revenue across the guest journey. The deal is backed by a strategic investment from Wavecrest Growth Partners, which is funding the merger and the combined platform's next phase of growth.

Franklin, a Danish payment card and finance platform for e-commerce merchants, raised €1.6M ($1.8M) in a seed round led by True Collective. The company's card targets high ad spenders with limits designed to keep campaigns running through peaks like Black Friday, with Franklin's AI collecting and reconciling receipts from ad platforms like Meta, Google, and TikTok to automate their bookkeeping. Franklin plans to use the capital to fund its push toward “agentic finance,” which is where AI agents reconcile and categorize transactions automatically rather than just showing numbers.

Reformation, a womenswear label that built a celebrity following among fans like Taylor Swift and Hailey Bieber, could go public as early as this summer, according to Wall Street Journal sources. Plans call for a confidential filing as early as next week and a listing by July, though sources stressed nothing is set and the timing may change. Reformation is on track for more than $500M in revenue this year, and its private-equity owner Permira, which bought a majority stake in 2019, would get a welcome exit through a listing by capitalizing on a hot IPO market.

Baseten, a platform that provides inference software and computing capacity for companies running open-source AI models, is finalizing a $1.5B funding round co-led by Altimeter Capital, Conviction, Spark Capital, Sands Capital, and Wellington Management. The round uses a dual-tier structure, with some investors buying in at an $11B valuation and others at $13B, nearly triple the $4B mark set by inference rival Fireworks AI last October. Baseten sits between enterprises and 20 different cloud providers, giving customers the software layer to run, optimize, and train open-source models on their own data at a fraction of closed-source cost.

Thanks for being a Shopifreak!

If you found this newsletter valuable, please leave a review on Google and share the newsletter with your friends and colleagues to help us grow.

See you next Monday,

PAUL

Paul E. Drecksler
🌐 Shopifreaks.com
🧑‍💼 Add me on LinkedIn
📧 [email protected]
📱 +1-828-273-3031
⭐ Leave A Review

PS: What do airlines serve passengers for breakfast? Plain bagels.