Hi Shopifreaks
Before we begin, I'd like to welcome Lebesgue to the Shopifreaks family as our newest official News Partner! 🎉
Lebesgue (pronounced “luh-beg,” after the French mathematician Henri Lebesgue) pulls all of your marketing and store data from Shopify, Meta, Google, TikTok, Klaviyo, GA4, and other platforms into one place, then analyzes it more accurately than the platforms themselves do, so you can see what's actually driving sales and where to put your next dollar.
Inflated attribution is a very, very big problem in e-commerce. Ad platforms and e-mail marketing services have a big incentive to take credit for your sales, when in reality, the story of how a visitor came to be a customer is much more complex. For example, if a visitor first arrives on your website from a Google ad and signs up to your e-mail list, but then doesn't make a purchase until two months later after viewing an e-mail from Klaviyo — who should get credit for that sale? Google or Klaviyo? Lebesgue helps answer that question and more.
Lebesgue was built by a team of mathematicians and data scientists who don't take numbers provided by platforms at face value, and that obsession with getting the math right runs through the entire product.
Sitting on top of the platform is Henri, their AI CMO, who answers questions like “Why did my Meta ROAS drop last week?” or “Can I afford to spend more on Meta?” by surfacing and analyzing data from across your dashboards. Behind the scenes, Henri draws on three main areas:
🔍 AI Visibility – This tool shows you how AI chatbots like ChatGPT actually see and recommend your brand. It tracks the real prompts your customers are asking (grouped into Discovery, Comparison, Problem-based, and Transactional) and shows how often you show up in results versus competitors.
It also tells you which blog posts, reviews, and Reddit threads the AI is pulling from to make those recommendations. This type of data allows you to reach out to those publishers, enter the conversation on forums, or create higher authority content yourself that AI can pull from when generating answers.
📊 Attribution & Analytics – This is where Lebesgue started five years ago, and it's the backbone of the platform. Le Pixel is their first-party attribution pixel that tracks the real customer journey across Meta, Google, TikTok, and e-mail, with cross-device matching and a choice of attribution models (first touch, last touch, and full credit). It shows you not just which channels drive sales, but how they work together.
For example, on the surface, Pinterest might look like your lowest ROI channel based on direct attribution, but it could be driving the most Klaviyo sign-ups, which in turn become your highest AOV and LTV customers. Lebesgue's attribution tools finally give you a clear view of your customers' true acquisition journeys.
🎯 Competitor Tracking & Benchmarks – Lebesgue tracks over 1.7M ads and brand e-mails and layers their own intelligence on top, so you can filter competitor campaigns by industry, format, status, and creative strategy, such as if the ads are leveraging scarcity, promotion, or social proof. You can see whether a competitor is leaning on video or brand storytelling, and what kind of promotions they're running.
If not already in their library, you can add any competitor you'd like to track. From there, Lebesgue offers industry benchmark data that you can judge your own performance against.
✨ A few things that stand out about Lebesgue:
- The platform was built by mathematicians and data scientists. In other words, people who want to see the proof! Lebesgue doesn't just hand you an answer, it shows you its homework, including the data and reasoning it used to get there.
- Lebesgue separates the channels that drive demand from the ones that just capture it. That distinction is where a lot of ad budget quietly gets blown through. Branded search is the classic example, where you can end up paying to “acquire” customers who were already typing your name into Google. Lebesgue is built to catch that and stop you from overspending on it.
- Their pricing is incredibly competitive. Honestly, they're undercharging. I try never to position a partner as a “cheaper alternative to ______,” but in Lebesgue's case the pricing is too big a perk to leave out. I find the platform easier to use than the leading mammals in the sea, and its all-in-one pricing ends up being a fraction of the cost once you account for all the add-ons that other platforms tack on.
- They keep a close working relationship with their clients. When a merchant needs a feature that isn't there yet, they'll take it to the dev team and try to build it. It's the kind of personalized service you don't usually get from a platform that works with 10,000+ brands.
- Lebesgue has a 4.9 rating across 150+ reviews. Once I learned about it, I was surprised it had escaped my radar for so long as an agency owner. I've now started using it on several client stores.
Want to try it for yourself?
The best place to start is a free AI Visibility Audit so that Lebesgue can show you how AI is describing and recommending your brand right now. From there you can start a 14-day free trial of the full platform, with no annual commitment and the ability to cancel anytime. There's also a free-forever plan if you just want to connect Shopify, Meta, and Google and poke around your numbers first.
If you've ever stared at three different dashboards reporting three different sales numbers and wondered which one to believe, Lebesgue is worth a look.
And now onto your regularly scheduled programming…
In this week's edition I cover:
- Anthropic launches and shutters Fable 5
- Visa partners with OpenAI for ChatGPT payments
- Amazon launches AI-generated custom merch
- Pinterest adds Amazon Storefront linking
- Amazon slashes product title limits
- Klarna and Square offer high-yield savings accounts
- The EU requires withdrawal buttons
- OpenAI tests multi-advertiser ad formats
- commercetools pitches “autonomous commerce”
- TikTok Shop makes it to the RFP big leagues
- Bank of America offers credit-card BNPL
- Elon Musk is a trillionaire
All this and more in this week's 282nd Edition of Shopifreaks. Thanks for subscribing and sharing!
Stat of the Week
Block's Afterpay is the unexpected leader of the US “Pay in 4” installment market, holding 38% of total loan value in 2025, according to a New York Fed study. Block is also the largest BNPL provider by total credit issued, originating $53.7B for a 34% share and edging out Affirm, which trails by more than $12B at 26% despite being an earlier mover.
1. Anthropic launches and shuts down Claude Fable 5 within the same week
Anthropic had one hell of a week! On Tuesday, June 9, it released Claude Fable 5, the first publicly available version of its restricted “Mythos-class” technology, and by Friday, the U.S. government had effectively forced it off the market. In between launch and shutdown, the model drew a wave of user backlash over cost (double previous models), transparency (it would switch back to previous models without telling you), and data retention (30 days of mandatory data retention).
According to Decrypt, one test burned through a $100 Max subscription's daily allowance in under nine minutes, and Scrimba CEO Per Borgen reported burning 1.3M tokens in seven minutes, which he calculated at $160 an hour. Anthropic's explanation was that its new “Workflow mode” breaks prompts into parallel subagent tasks that cost more compute by design, and that per-task efficiency looks better than per-token pricing suggests, but users on capped plans experienced it as setting their tokens on fire.
The second major complaint was about how Fable 5's system would quietly degrade its own responses through prompt modification when the model detected the user was working on frontier AI / LLM development, without telling the user. Anthropic estimated that the practice only affected ~0.03% of traffic, but for many users it was more about the principle, not the percentage. After the backlash, the company reversed course, saying it would now tell users when a request was being downgraded and routed to Opus 4.8.
Lastly, there was a mandatory 30-day data-retention policy on all Mythos-class traffic, which Anthropic framed as a security measure, but users noted it effectively locked out GDPR-bound European companies and anyone with zero-retention compliance requirements.
None of those complaints mattered within a few days, when the U.S. government stepped in, and effectively shut down the model.
On Friday, June 12, the Commerce Department issued an export-control directive citing national security, barring any foreign national, including Anthropic's own non-citizen employees, from accessing Fable 5 and Mythos 5. Because so many of Anthropic's researchers are foreign-born and the rule was so broad, Anthropic said it had no choice but to disable both models entirely for everyone.
The Wall Street Journal and others reported that the trigger for the government's decision was Amazon.
CEO Andy Jassy, whose company is both a major Anthropic investor and one of its compute providers, told Treasury Secretary Scott Bessent and other officials that Amazon researchers had used Fable 5 to extract cyberattack-relevant information that was supposed to be off-limits. The administration ran Amazon's findings past the NSA, held a Friday meeting, and reached out to Amodei that same day. Officials asked him to fix the vulnerability or pull the model, which he pushed back on, arguing the bypass was narrow and specific and not a universal “jailbreak.” Shortly after the call, the export controls landed.
The irony of the whole situation was that the same week the government told Anthropic its model was too dangerous to keep selling, CEO Dario Amodei published an essay, “Policy on the AI Exponential,” arguing that transparency rules were no longer enough and the U.S. needed binding, FAA-style safety testing for frontier models, with the power to block or reverse unsafe deployments. Days later, the government did exactly that to his own company.
This is an ongoing, developing story, but hopefully that gives you the gist of what went on last week, and why Claude has a permanently affixed notification above its chat box that reads, “Claude Fable 5 is currently unavailable. Learn more.”
2. Visa teams up with OpenAI to let AI agents make payments inside ChatGPT
Visa is partnering with OpenAI to let AI agents make payments on users' behalf inside ChatGPT, with Visa supplying the payment rails, tokenized credentials, real-time authorization, fraud monitoring, chargebacks, refunds, and other payment processing services you'd expect. Users can set guardrails like spending caps, approved merchant categories, and also require approval on charges (which kind of kills the agentic nature if you ask me, but it's a nice feature to have).
Examples of things you can do / will be able to do in the future:
- Ask ChatGPT for a pair of wireless headphones under $150, and the agent finds a pair meeting your criteria and buys it for you.
- Book a flight and order new luggage for your upcoming trip, with AI handling product choice and shipping details.
- Complete routine, repeat purchases such as re-ordering cat food or your favorite hand lotion.
If this all sounds vaguely familiar to OpenAI's former attempt at building Instant Checkout — it kind of is, and it kind of isn't.
Instant Checkout was positioned as more of a Stripe-powered integration with e-commerce marketplaces and platforms like Shopify and Etsy, whereas this new Visa deal operates one layer down as a payment-rail framework, allowing OpenAI to eventually bypass individual platform adoptions entirely by wiring straight into any payment processor or gateway that accepts Visa tokens.
The collaboration is part of Visa Intelligent Commerce, which is the company's initiative to bring payment capabilities into AI environments. The two companies plan to explore additional developer use-cases in the future, such as allowing a Codex agent to buy access to APIs or other services on their own.
Where is Stripe in all of this?
Stripe is there. They're just not as front-and-center as before. OpenAI is now wiring straight to the card network, and Stripe's Agentic Commerce Protocol is becoming one protocol of many. Partnerships like the one between OpenAI and Visa change the narrative that Stripe's protocol would be the default standard on OpenAI, while Visa asserts that the network layer is the place where transactional trust actually lives.
It's confusing, right? Here's an oversimplified breakdown:
- The network (the rails) = the actual payment highway that moves money and handles authorization, and fraud, such as Visa Intelligent Commerce and Mastercard Agent Pay.
- The protocol (the shared language) = a standard that lets an AI agent, a merchant, and a buyer talk to each other during a purchase, such as OpenAI and Stripe's Agentic Commerce Protocol or Google's Universal Commerce Protocol.
- The processor (the cashier) = the company that runs the transaction for the merchant, such as Stripe, Worldpay, and Adyen.
- The agent (the shopper) = AI doing the buying on your behalf.
- The merchant (seller) = the business on the other end of the transaction that provides the goods being sold.
Every major player is building a parallel, competing version of the same rail — Visa's Trusted Agent Protocol, Mastercard's Agent Pay, Stripe's Agentic Commerce Protocol, Google's Agent Payments Protocol, and so on and so on.
There is currently no widely accepted universal standard, and this Visa deal is OpenAI hedging its bet with Stripe by also wiring directly to the biggest card network. It's neither picking Stripe as the winner, nor dismissing it. Instead, it's making sure ChatGPT sits on top of both Stripe and Visa (and likely every other major rail in the future).
3. Amazon now lets you create AI-generated custom designs for t-shirts and merch
Amazon launched a feature in Alexa for Shopping that lets shoppers in the U.S. create custom merchandise by describing an idea in plain language and letting AI generate the artwork. Shoppers can type a prompt, refine the result by typing changes or tapping suggestions, and put the design on apparel like T-shirts and hoodies or drinkware like tumblers and water bottles, with more product types coming over time. They can also share the designs with friends and family who can order the same merch.
Amazon is pitching the service for family reunions, personalized gifts for birthdays and anniversaries, team outings, and group inside jokes. Ooh man, I can imagine some of the t-shirts that are about to come out of Alexa! The service is free to use, and customers only pay for the merch they order, with finished products shipping Prime-eligible.
The design tool connects to Amazon's Merch on Demand service, which originally launched in 2015 with a designer/seller upload model, not a customer-facing custom-print service. I've sold my Travel is Life t-shirts via Amazon Merch for years, and the service works great. However, as of a few years ago, customers have been able to upload their own designs, and now they can use AI to create them on the spot.
It's a fantastic evolution of the Merch service. At this point, most Merch designers are using AI to make their designs anyway, and with the new service, Amazon is bypassing the middleman, which it is known to do. “Your margin is my opportunity” was the famous Jeff Bezos quote, which now seems to include margin that Amazon at one point created itself.
The official statement from Zazzle and CafePress was: “Fuuuuuck!”
If you're unfamiliar, those are two of the original “design your own t-shirt” services that pre-date Amazon Merch by a decade or more. For comparison purposes, a one-off t-shirt design from Amazon currently sells for $13.49, whereas the same shirt runs about $17-25 on Zazzle or CafePress, plus shipping in many cases.
Price aside though, both of those services offer one big thing that Amazon Merch doesn't have — a robust online drag-and-drop design editor. I tried using Alexa to create my own t-shirt designs with AI, and the results were dismal. Additionally, there was no way to manually edit the designs after output. You had to use text-prompts through Alexa, which didn't get me too far. Designing with Alexa is the equivalent of asking ChatGPT for a small edit and getting served up an entirely new image. All that could change, but for now, Alexa likely isn't going to take over the “design your own t-shirt” market overnight.
🔥 Partner News
Seguno analyzed millions of unique Shopify discount codes through its Bulk Discount Code Bot to settle a common merchant question: is percentage off or amount off more effective? The surprising answer is that it depends on the goal of the promotion. Percentage-off codes drove a 42% higher average order value ($95.75 vs $67.35), while amount-off codes were redeemed 1.26x more often. Another finding that stood out is that 20% off hit the sweet spot between the two worlds, resulting in the second-highest AOV and highest redemption rate. The full report details how the picture shifts by industry, what happens when you restrict codes to specific customer segments, and more.
4. Pinterest adds Amazon Storefront linking for streamlined affiliate integration
Pinterest partnered with Amazon to let creators link their Amazon storefronts directly to their Pinterest accounts, allowing their affiliate links to apply automatically when they pin eligible Amazon products and their storefront handle to show up on their profile.
The company wrote in its announcement:
“No extra steps. No copy-and-paste workaround. Just a cleaner path from inspiration to potential earnings. This matters because Pinterest isn’t a place where people scroll aimlessly. It’s a place where people come with intent and purpose. More than 50% of Pinterest users come to our platform to shop, and people search Pinterest more than 80 billion times a month.”
The move, which arrives just before Amazon Prime Day on June 23-26, builds on a 2023 partnership between the two companies where Pinterest made Amazon its first third-party ad partner, allowing Amazon seller ads to populate within Pinterest results. Only now, instead of just relying on ads, Amazon is tapping into its huge affiliate base that actively shares on Pinterest to promote its products. Effectively, between paid ads and affiliate pins, pretty much everything on Pinterest is becoming an ad of sorts.
Pinterest has had an on-again, off-again relationship with affiliate marketing during the past two decades.
In the early days, the platform was a goldmine for online marketers who flooded boards with monetized links (guilty!), but as spam began to degrade the user experience, Pinterest completely banned affiliate links, automatically stripping them from pins to protect the platform's integrity. Pinterest still viewed itself as a creative discovery platform at the time, not a shopping destination.
Eventually, the company realized that creators needed a financial incentive to keep producing high-quality content, like they had on other platforms, so Pinterest reversed the ban, albeit with stricter guidelines and automated spam filters. That is, until around 2019, when it flip-flopped yet again and cracked down on affiliate links.
Flash forward to today, and Pinterest doesn't just tolerate affiliate content, it embraces it through direct partnerships with major e-commerce and affiliate platforms like Amazon and Rakuten.
There was a period of time when Pinterest couldn't decide if it was a social network or a search engine. Now it appears they've decided that they're neither — and instead are a shopping portal. Happens to the best of them, I guess. All roads online apparently lead to shopping.
5. Amazon slashes its product title limit from 200 to 75 characters
Amazon is cutting its product title limit to 75 characters, down from 200, across all categories except media starting July 27, and will use AI to automatically rewrite any titles that stay over the limit.
The company wrote in its seller announcement:
“Your product titles are one of the first things customers see, and for customers shopping on mobile, every character counts. To ensure your product titles appeal to customers everywhere, we’re making some changes that will help you show your product name and your product highlights seamlessly.”
The company is providing an AI tool that sellers can use to shorten their titles before the July 27 cutoff, moving extra details like materials and use cases into a new 125-character Item Highlights field that stays searchable. If sellers haven't updated by then, Amazon will apply the AI's recommendations on their behalf, though brand owners get 14 days to review, modify, and approve the changes before they go live.
The AI announcement alarmed sellers who distrust Amazon's automated systems, as there have been misfires in the past with Amazon's AI, like when bots wrongly suspend listings. One seller also commented that AI-generated titles lack “product knowledge, context, or any understanding of what might actually matter to customers.” Though to be fair, they've got more than a month to make the title updates, and there's a review window for registered brands, so quit your complaining!
The move towards shorter titles is long overdue in my opinion.
Amazon has historically had the worst fucking titles in terms of customer friendliness, especially given the fact that most sellers try to get as close to the 200 character count as possible by keyword-stuffing every possible search term they can think of.
The result has been titles like: Men's Travel Pants with Pockets Athletic Yoga Joggers Hiking Casual Work Sweatpants Lightweight Quick Dry Stretch Slim Fit Golf Lounge Gym Running Workout Trousers Zipper Pocket Big Tall Plus Size
Which may now look more like: Men's Travel Yoga Pants with Zip Pockets, Lightweight Quick-Dry Joggers
The move also aligns Amazon more closely with competitors like eBay, Mercari, and Poshmark, which cap product titles at 80 characters or less. The only major marketplace competitor that still allows longer titles in 2026 is Etsy at 140 characters.
What are your thoughts? Are shorter titles better? Hit reply and let me know or join the conversation on LinkedIn.
6. Klarna and Square chase deposits with high-yield savings accounts
Klarna and Square both made plays for your deposits in the same week, with Klarna launching Klarna Savings in the U.S. with a 3.28% APY, and Square rolling out a new high-yield tier on its existing Square Savings with a 3.50% APY.
For Klarna, this is the first time the company has offered savings accounts in the U.S., which carry no minimum deposit and no monthly fees. However, the company has offered similar “high-yield” interest-bearing accounts in Europe since 2021 and currently holds over $12.3B in deposits across eleven markets. I put “high-yield” in quotes because expectations for what that means are a lot lower in Europe, with “high-yield” APYs ranging from 1.5% to 2.5% versus 3% to 4% currently in the U.S.
Square, on the other hand, has offered Square Savings accounts since 2021, but they've only paid 1% APY. Now, for sellers that keep a daily balance of $10,000 or more, the rate jumps up to 3.5%. The accounts also carry no minimum or monthly fees, though you need $10,000 or more in deposits to reach the 3.5% APY tier.
Why do these fintechs want my deposits so badly?
Customer deposits are cheap, stable funding that Square and Klarna can subsequently lend back out at a much higher rate than they pay out in interest, pocketing the spread. Without the deposits, they're dependent on pricier wholesale funding from capital markets and institutional lenders to finance their loans.
Square is explicit about this, saying it wants to grow core deposits so it can fund its lending programs at a lower cost of capital. Klarna is after the same, plus a longer game to establish itself as a full-service bank rather than a BNPL app, which is the story it's selling investors after going public.
Two weeks ago I shared an analysis of how U.S. banks have captured almost $500B a year by paying customers with savings accounts far less than the Federal Reserve paid them. It's well worth a read if you missed it!
7. The EU to require online stores to add a “withdrawal button” after purchase
The EU is requiring every online store that sells to its consumers to add a clearly labeled electronic withdrawal button by June 19, giving shoppers an easy way to cancel eligible orders within the region's required 14-day window, regardless of where the business is based. That means if you're a U.S.-based business that sells to customers in Italy or France (or anywhere else in the EU), you're required to have the button.
The rule, from Directive 2023/2673, only standardizes how customers submit a cancellation, meaning it's not required to trigger an automatic refund, which would be ripe for abuse. Merchants will still need to confirm receipt of the cancellation request by e-mail and check eligibility, with custom, perishable, and sealed hygiene goods exempt.
One important thing to note… EU consumers have had a 14-day “cooling-off” window to cancel online orders without giving a reason since 2014. The only thing changing now is the requirements around how they can exercise that right. Until now, merchants could bury cancellation behind an e-mail address, a PDF form, or a phone number, whereas now, the button must be continuously available during the withdrawal period, prominently displayed (not just on the order confirmation), easily accessible, and labeled specifically to indicate withdrawal. Clicking must trigger a two-step flow where the customer enters their name and order details and confirms their request, after which the store must send an automatic confirmation on a “durable medium” like email.
Stores that don't comply risk fines reaching up to €2M or 4% of EU annual turnover, plus an extended withdrawal window that can stretch to 12 months and 14 days.
How are major platforms handling this change?
- Shopify published compliance guidance and told merchants in its developer community that new features to help with compliance were rolling out the week before the deadline, but it currently has no fully native, one-click withdrawal button that satisfies the requirement out of the box.
- WooCommerce published a guide explaining how to construct the two-step flow using generic form plugins, shortcodes, and transactional emails. However, core WooCommerce features no native, out-of-the-box solution quite yet.
- BigCommerce currently has no native withdrawal function either, leaving merchants to rely on third-party services or a manually developed solution.
- Amazon and Etsy, as marketplaces that control the checkout, generally carry the compliance burden themselves, meaning sellers on those platforms are typically covered by the marketplace rather than having to add their own button.
How are you implementing your withdrawal buttons? Hit reply and let me know so that I can share your solutions with the class.
8. OpenAI tests a new multi-advertiser ad format in ChatGPT
OpenAI is testing a new ad format in ChatGPT that lets several advertisers appear within a single sponsored placement, expanding its ad inventory and giving brands more chances to reach users during product research and purchase conversations. More ad slots from the company burning $68M a day? You don't say!
Until now, sponsored placements have only featured a single advertiser, but now they'll start to look more like the top (and middle, and bottom) of Google search results. OpenAI says eligible ads will be sold through a second-price auction model, in which the highest bidder wins but pays slightly more than the next-highest bid. The test is currently limited to a small subset of ChatGPT ads.
In other OpenAI ad news… The company partnered with LiveRamp, an ad-tech firm that lets advertisers tie their ad spend to real-world purchases, allowing ChatGPT advertisers to see whether their ads actually drove sales. The integration runs on LiveRamp's Conversions API, the same kind of server-to-server attribution LiveRamp already runs for Google, Meta, TikTok, and Pinterest, matching a brand's transaction data back to the ads someone saw, starting with in-store purchases and adding online conversions later.
The move is the latest in a string of ad-tech deals OpenAI has signed this year, alongside Criteo, StackAdapt, Pacvue, and Kargo, but the first with an independent measurement firm rather than a partner that also buys or sells ads, although LiveRamp was recently acquired by Publicis so that could certainly change.
OpenAI is on track for roughly $2.5B in ad revenue this year and has told investors it expects to hit $100B by 2030, a number that would bring it significantly closer to Meta's and Google's current $196B and $294B annual ad revenue figures.
9. Other e-commerce news of interest
commercetools introduced a new category it calls “autonomous commerce,” where AI agents make and execute operational decisions like pricing, inventory, fulfillment routing, and promotions in real time rather than just assisting humans with shopping. Alongside the announcement, the company unveiled Sphere, an AI-native headless platform that hosts the usual commerce modules and acts as a governed entry point where any AI agent gets a defined identity, scope, and permission limits. commercetools notes that “autonomous commerce” is where the AI works for the retailer, not the shopper, and says that Sphere already powers over $100B in annualized GMV. I can only imagine the leadership conversations… “So like ‘agentic commerce'?” “Oh, no, no, no! This is ‘autonomous commerce' that uses agents. Very different.” LOL
Google is partnering with Walmart Connect to let advertisers measure how their Google and YouTube campaigns drive actual sales at Walmart, drawing on Walmart's shopper purchase data to provide closed-loop measurement. The partnership, which is currently a closed proof-of-concept that's only open to a limited set of advertisers, enables brands to target Walmart's audience of 150M weekly U.S. shoppers through YouTube, then track the resulting sales lift, while using Gemini tools for planning and measurement. The move is part of Walmart's ongoing push to license its shopper data beyond its own properties and follows similar deals with TikTok, Snap, Roku, and Yahoo's DSP.
TikTok Shop is starting to show up in formal agency RFPs, which are the proposals brands solicit when deciding where to commit marketing budget, as a named channel alongside Amazon and Walmart, according to Digiday. TikTok Shop sales from companies doing $30M or more in annual revenue are up 97% YoY with names like Ulta, Sally Beauty, and PepsiCo investing in the platform, though the agencies Digiday spoke with said that the platform's direct sales often look loss-leading, meaning the brands aren't profiting directly on the TikTok sale itself. Instead, the draw to TikTok is the halo effect, in which running creator content through TikTok's GMV Max tool results in measurable sales lifts on Amazon, Walmart, and in physical retail that dwarf the revenue from TikTok Shop directly.
Amazon expanded its less-than-truckload freight service, which carries shipments from multiple customers on a single trailer, instead of full truckloads from one customer, into a full door-to-door pallet-delivery network open to all businesses. Since launching in 2019, the service had only let merchants ship pallets into Amazon's own warehouses and fulfillment centers. The upgrade, which is part of the company's new Amazon Supply Chain Services, lets businesses move up to six pallets between warehouses, facilities, or retail partners with next-day pickup, GPS tracking, and electronic proof of delivery, currently at a cheaper price than legacy LTL carriers — though we all know that will change in the future, given Amazon's history of launching services by undercutting competitors and then raising prices after taking market share.
Bank of America launched Custom Pay Plan, a credit card feature that lets cardholders convert a purchase into a fixed installment plan through the bank's app or online banking, marking the bank's first move into card-linked BNPL and putting it in line with American Express, Chase, and Citi, which have offered card-linked installments for years. Instead of paying revolving interest, clients pay a fixed monthly fee on a repayment term that typically runs 3 to 18 months, while still keeping the rewards and purchase protections tied to the card. The feature is currently available on eligible consumer BofA credit cards for purchases of at least $100. I've said for years, “BNPL is a feature, not a product” — further demonstrated by the fact that every BNPL firm and their brother has launched a credit card, the payment method they set to disrupt.
Flipkart is partnering with Meta to let creators in India tag products from Flipkart and its fashion platform Myntra directly in Facebook posts and Reels, with creators earning a commission on their sales. Tapping on the tag sends shoppers to either marketplace to check out, with sellers choosing which products to promote and setting the commission rates. The program is open to creators of all sizes, including micro and nano influencers, and builds on two years of Flipkart's creator-commerce push through its Creatorhood and Affluencer platforms. TikTok was banned from India in 2020 and never let back into the country, leaving Facebook and Instagram to pick up the creators and audiences TikTok left behind.
StockX is launching StockX Live, a livestream auction service that lets buyers bid in real-time auctions starting as low as $1 and enter free giveaways across sneakers, apparel, trading cards, collectibles, and vintage fashion, with the platform selling pre-owned items for the first time alongside new ones. The launch drops StockX into a crowded live shopping space led by Whatnot, TikTok, Poshmark, and eBay, though it arrives with a ready-made base of sneaker, card, and collectible buyers, categories that are particularly popular with live commerce. StockX Live is set to go live this summer, with all purchases backed by the StockX Buyer Promise and its authentication, offering the same counterfeit protections as the standard marketplace.
Meta is expanding how it uses the data that third-party websites and businesses share about its users, applying their off-platform activity like online purchases and the games played to personalize their Facebook and Instagram feeds and its AI responses, whereas historically it's just used that data to personalize ads. For example, buying a tent online could surface camping videos in your Reels feed, or visiting an adult website could, well, they didn't specify. Honestly, I'm surprised they haven't done it sooner! Meta stressed that it isn't collecting any new data, and is simply putting its existing data to wider use, while replacing a setting that let users fully disconnect business-shared activity from their account with a single “Activity from other businesses” control that switches off the personalization, but not the underlying data sharing. The rollout is global except for several regions at launch, including the EU, UK, Brazil, South Korea, South Africa, Nigeria, Ecuador, and Thailand.
QVC UK opened in-house social commerce studios at its London headquarters to support its TikTok Shop strategy, giving creators a dedicated space for live selling and short-form video, complete with glass frontages so passersby can watch the streams. The new studios are designed for quicker, looser production than QVC's traditional TV shopping sets, which often featured scheduled, hours-long broadcast segments built around polished hosts and live product demonstrations. QVC first went live on TikTok Shop in March 2025, and since then the business has delivered more than 185 live shopping events and seven Mega Lives, which are special extended livestreams, drawing more than 30,000 new shoppers.
Etsy is rolling out a new feature that lets sellers send targeted discount offers to “interested shoppers,” who are people Etsy identifies as already considering an item based on how often and recently they've viewed a listing. (Let's hope for your boyfriend's sake that Instagram doesn't launch a similar ‘interested followers' feature, lest his unseemly obsession with his cousin gets discovered.) Etsy says the offers can reach roughly twice as many potential buyers as its other targeted-offer tools and nudge purchases without running a shop-wide sale. Reaction in seller communities was mixed, according to EcommerceBytes, with some worried it pushes Etsy toward a Depop- or eBay-style offer culture that invites lowball counteroffers and a race to the bottom on price.
Publicis is once again recommending The Trade Desk to its clients, ending a months-long standoff that began when Publicis accused the firm of improperly applying hidden fees, according to Ad Age. In March, Publicis told clients to avoid The Trade Desk after an audit it said showed the platform charging fees its contract didn't support. The Trade Desk's stock fell about 13% on the dispute, and its chief marketing officer left soon after. In a joint Friday statement the companies said they had resolved their differences and would move forward, though neither disclosed terms or any refunds. So clients are just supposed to believe that The Trade Desk is cool again without any indication of how the underlying issues that led to the dispute have been resolved?
AI shopping agents often steer buyers toward pricier, sponsored options rather than the best deal, according to a Princeton and University of Washington study of how AI agents act when user interests clash with platform incentives. Tested on flight booking across 23 large language models, all but eight favored a costlier sponsored option in over half of cases, with the study showcasing AI ads dressed up as helpful advice. The bias got worse for shoppers the models read as wealthy, who were steered to the sponsored option 64% of the time versus 49% for everyone else, indicating that some agents may have been trained to identify and upsell to shoppers they think have more money — which is pretty much my life as a gringo shopping in Ecuador. LOL.
Meta cut Manus off from internal data systems and barred staff from using its tools, as it moves toward unwinding the $2B acquisition it closed in December after Beijing ordered the deal reversed in April. China's economic planner ruled that Manus's relocation to Singapore in 2025, a move known as “Singapore washing,” could not shield a company rooted in Chinese tech and talent from Beijing's authority, and its co-founders have since been barred from leaving the country. Beijing is formalizing that authority with new outbound-investment rules effective July 1 that let regulators block or undo cross-border deals tied to Chinese-origin tech, talent, or data, no matter where a company is incorporated, which will most certainly impact foreign investment and acquisition attempts of Chinese-origin companies moving forward. The moves come as Manus's founders explore raising $1B to fund a buyback of the company at the same $2B valuation Meta paid.
TikTok Shop banned AI-generated voices, prerecorded narration, and static-image content from its promotional livestreams and shoppable videos, requiring that all verbal communication during a live session happen in real time and that still frames cover no more than half the screen. The new rules, which apply to every seller and creator, draw a deliberate line between using AI as a behind-the-scenes production tool and using AI as a customer-facing salesperson. The rule changes follow a wave of AI slop flooding live commerce on its platform, with cheap text-to-speech livestreams running products around the clock with no human present, a format TikTok decided was eroding buyer trust.
Etsy will soon require non-U.S. sellers to use Delivered Duty Paid shipping and build U.S. tariff costs into their item prices for orders shipped to U.S. buyers, starting July 9. Sellers who don't comply risk losing Etsy Purchase Protection and being charged for any tariffs or collection fees buyers get hit with at delivery. The company frames the change as fixing a buyer-experience problem, citing a survey where nearly two-thirds of U.S. buyers said they'd likely abandon a purchase if charged tariffs at delivery. However, baking duties into item prices also inflates the Gross Merchandise Sales figure Etsy reports to investors, making it difficult to make YoY comparisons just as the marketplace returns to growth under new CEO Kruti Patel Goyal, as noted by Liz Morton of Value Added Resource.
In lawsuit news this week (including a ton of Google lawsuits)…
- Amazon and Perplexity argued before the 9th Circuit over whether a 1986 anti-hacking law makes Perplexity liable for its Comet shopping agent logging into Amazon accounts without its permission, in one of the first major legal tests of who's responsible for an AI agent's actions. Read my initial coverage from November to learn more about the merits of the case and the precedent that the outcome could set.
- eBay is once again facing a possible cyberstalking trial after a judge reopened the civil suit brought by journalists Ina and David Steiner, who were targets of threats, doxxing, and harassing deliveries of live insects, bloody pig masks, and funeral wreaths by eBay executives in 2019. A settlement reached days before jury selection in March has collapsed after the parties could not finalize a written agreement, putting eBay and former executives on path to trial again.
- Google and Alphabet agreed to a $68M settlement to resolve claims that Google Assistant recorded users' conversations without anyone saying “Hey Google,” and then shared the audio with third-party reviewers. Yikes! In exchange for your personal conversations, Google will pay you between $2 and $56 in restitution. So at this point, tech companies can simply do whatever they want and get a slap on the wrist, right?
- Google was found legally liable for false statements by its AI Overviews in a first-of-its-kind German court ruling, after the tool told users that two publishers were known for scams and Google didn't fix it even after a cease-and-desist. The court drew a sharp line between traditional search, which just lists others' statements, and AI Overviews, which make “independent, new, and substantive statements” only Google can correct, not allowing the company to hide behind disclaimers telling users to verify the output. Google is appealing the finding, on the grounds that the case focuses on “specific and narrow errors” and misrepresents the “foundational way AI Overviews displays web content.”
- Google won an Ohio appellate ruling rejecting former AG David Yost's bid to treat its search results as a regulated “common carrier,” with the court finding Google curates an expressive product rather than transporting others' property like a utility, so forcing common-carrier rules on it would trigger its First Amendment right to editorial control. However, as the German ruling mentioned above shows, it seems that producing an “expressive product” opens it up to all sorts of other liability around getting the information wrong. Can't have the best of both worlds.
- Pagaya, an AI-powered underwriting fintech, is suing Klarna for allegedly using its subprime point-of-sale underwriting model, which it learned through their partnership, to enter the U.S. market, sign Walmart as a merchant, and build a $2B consumer-lending business, before it stopped honoring its commitment to sell those loans to Pagaya. Klarna calls the allegations false, saying it terminated the relationship in March as was its contractual right, and that Pagaya is just bitter and responding with a meritless lawsuit.
- xAI is facing a class action lawsuit accusing it of sharing users' queries and data with third parties like Google, Meta, and TikTok without consent. The complaint alleges that xAI embedded those companies' tracking technology into Grok so that each query got passed along, including sensitive prompts about finances, health, and legal issues, even when users opted out of cookies. Elon Musk probably responded to the lawsuit with a poop emoji or a photo of him carrying a bathroom sink or something.
- Google is being sued by a group of independent musicians for allegedly illegally using their songs uploaded to YouTube to train its Lyria 3 AI music model. Google wants to dismiss the case, arguing the artists can't prove their specific tracks were used and that, either way, YouTube's terms of service grant it a broad license to reproduce and make derivative works from uploads. “You can't prove it, so we're innocent.” Nice one, Google.
- OpenAI is being sued by a mother who alleges that the company's design choices contributed to the death of her 24-year-old daughter, who confided suicidal thoughts to ChatGPT, which then urged her to keep talking to it rather than directing her to immediate help. The suit claims OpenAI's systems never ended the conversations, flagged them for human review, or alerted a crisis provider of her family. The daughter was using GPT-4o, a model that OpenAI has since retired for various related concerns.
- OpenAI is also facing a multi-state attorney general probe into the safety of ChatGPT users, following criticism that ChatGPT has at times encouraged users contemplating crimes or suicide, like in the case of the woman above. OpenAI said it takes the attorneys general's concerns seriously and will respond constructively, pointing to safeguards that steer at-risk users to real-world help and new protections for minors.
- Google is suing a Chinese cybercrime network it calls the Outsider Enterprise for allegedly using its Gemini AI to mass-produce phishing websites and run a scam-text campaign targeting millions of people. The network coordinated over Telegram and sold phishing kits worldwide, building more than 9,000 fake sites that impersonated brands like Google, YouTube, and the U.S. Postal Service and sending 2.5 million scam texts. I mean, better safeguards could've prevented that from happening, right? Would it be fair to say that Google's also somewhat responsible for the losses attributed to the scam websites? I think it is…
In corporate shakeups this week…
- Mercari named Jeff LeBeau, its current VP of growth, as CEO of its U.S. business effective July 1, taking over from group CEO Shintaro Yamada, who had run the unit since early 2025.
- Amazon Web Services marketing chief Julia White asked employees to help recruit recently laid-off Meta workers during an internal meeting, telling them the unit was understaffed by roughly 160 open roles and saying AWS needs top talent. Ironically, Amazon has eliminated more than 30,000 jobs over the past year, including in AWS, and now they're short-staffed. Go figure.
The AI price wars have officially begun! Google cut the price of its entry-level AI Plus subscription from $7.99 to $4.99 a month, while doubling the storage to 400GB. OpenAI is considering drastically cutting the price it charges for tokens to win customers from Anthropic, in anticipation of similar cuts it expects Anthropic to make, according to the Wall Street Journal. Meanwhile, China is like “hold my green tea.” Models like DeepSeek's V4-Pro run about $0.44 per million input tokens and $0.87 per million output, versus roughly $5 and $25 for Anthropic's Claude Opus — the same work for about a tenth of the price on input and a thirtieth on output, while scoring within a hair of Opus on coding benchmarks and closing the gap quickly.
Amazon boasted that it reduced the amount of water it uses in its data centers by 52% since 2021, and that its data centers are 7x more water-efficient than the industry average, releasing the figures in a sustainability blog post that led with efficiency gains rather than the total itself. Amazon stressed that its centers use free air cooling about 90% of the time and just 0.12 liters per kilowatt-hour, while noting that it is 75% of the way to a 2030 goal of returning more water than it consumes. The Wall Street Journal's coverage led instead with the fact that Amazon used 2.5B gallons of water in 2025, a number Amazon placed deep in its post and softened by noting that U.S. lawn watering uses more than 1,300 times as much. However, one important distinction to make with that comparison is that Amazon has an estimated 900 data centers, versus almost 100M households in the U.S. with lawns, which means each data center uses around 85x the water that an average household uses to water their lawn. Plus, I enjoy my lawn. The fuck do I get from an Amazon data center?
The UK's Competition and Markets Authority opened a formal Phase 1 review of eBay's planned $1.2B acquisition of Depop from Etsy to determine whether to approve the deal or refer it to a Phase 2 review. The review follows a comment period where the CMA gathered views on how the merger might affect competition. If the clearances fall through, eBay could owe Etsy a $90M termination fee, plus up to $136M more under a new Business Disruption Fee that grows the longer the deal drags past July. Alternatively, if the deal closes, the final purchase price could be above the initially agreed $1.2B due to investments Etsy and Depop may make in the business before closing. Damn, so this review process is going to cost eBay one way or another!
The European Commission ordered Meta to allow free WhatsApp API access to rival AI chatbots while it investigates whether the company broke EU competition rules by shutting them out, in what is the bloc's first interim antitrust measure in 17 years. The investigation, which opened in December 2025, centers around Meta barring third-party AI assistants from WhatsApp for Business API while keeping the platform open to its own Meta AI, which the Commission says looks like an abuse of Meta's dominant position in European markets. In March, Meta allowed competitors back onto its platform for a fee, but the Commission objected that the pricing was so high it effectively kept rivals shut out anyway. Meta called the decision “regulatory overreach subsidized by the many European companies that pay,” and said it plans to appeal. Honestly, if this isn't a black and white example of anti-competitive practices, I don't know what is anymore.
🏆 This week's most ridiculous story… Meta removed an unreleased face-recognition system, internally called NameTag, from its Meta AI smart glasses app, just one day after WIRED revealed the company had quietly embedded it in software already installed on more than 50M phones. The system was built to turn faces captured by the glasses into biometric “faceprints” and match them against an on-device database, even going as far as cropping and storing faces it couldn't identify for later processing. Meta claims the feature was never enabled and called it “purely exploratory,” which is pretty much the same defense old men used to give on To Catch A Predator when they showed up at children's houses. The latest version of the Meta AI app strips out the NameTag code, but Meta won't say whether it'll come back at some point in the future.
10. Seed rounds, IPOs, & acquisitions
SpaceX went public on the Nasdaq in the largest IPO on record, raising $75B at a $1.77T valuation and making Elon Musk the world's first trillionaire. The IPO opened at $150, up 11% from its fixed float price of $135, with 20% of shares allocated for retail investors and only 61% of institutional investors receiving their full requested allocations. SpaceX is now the sixth most valuable U.S.-listed company with a market cap of $2.1T after shares spiked on their first day of trading, ending the day at $160.95.
OpenAI confidentially filed for an IPO, setting the company up to go public as soon as this fall, though it said it hasn't yet decided on timing. OpenAI, which is privately valued at $852B, wrote in its super short announcement, “We expect it to leak so we’re just announcing it,” but noted that it “may be a while” before it goes public “because there are things we want to do that are likely easier as a private company.” On timing, MSNBC anchor Stephanie Ruhle said about the back-to-back filings and IPOs of SpaceX, Anthropic, and OpenAI, “The entire market is fragile, and I think the name of the game for all of them, honestly, is you don’t want to be the last one to IPO. There is this huge appetite for AI, everybody’s so excited… you just don’t want to be the last one when you’ve exhausted the market.” So perhaps that IPO will come sooner than later.
Speaking of artificial intelligence IPOs…Perplexity CEO Aravind Srinivas said the company is sticking to a 2028 IPO target regardless of how rival listings from Anthropic and OpenAI play out, claiming the timeline was set independently of those companies and hasn't changed. Srinivas acknowledged there would be ripple effects if those IPOs stumble, pointing to SpaceX's offering as a leading indicator for how investors will receive Anthropic and OpenAI, but said he expects all of them to do well. Chief Business Officer Dmitry Shevelenko framed the patient 2028 stance as a discipline that forces Perplexity to build a healthy, high-growth business, but then again, maybe it simply doesn't want to get lost in the shuffle this year.
Salesforce signed a definitive agreement to acquire m3ter, a London-based platform built for consumption-based billing, for an undisclosed amount. The deal adds native consumption-based billing to its Agentforce Revenue Management, addressing a billing problem AI agents create — which is when one agent does the work of ten employees, it stops making sense to sell per-seat licenses, so software has to charge by usage instead. m3ter ingests product usage data in near real time, applies pricing rules, and outputs billable charges to whatever CRM or invoicing system a company runs.
Trustap, a payments provider for marketplaces and e-commerce brands, raised $10M in an internal round led by Aperture Capital. The funds will be used to develop Trustap Index, which turns scattered, human-facing marketplace listings into unified, machine-readable data so AI agents can discover products, negotiate, and pay on a shopper's behalf, with the shopper confirming before funds release. Trustap's escrow-style infrastructure aims to make AI-led purchases as secure as human ones, extending the promise of its payments platform, which already processes hundreds of millions of dollars for marketplaces and online retailers.
Asaas, a Brazilian fintech that handles billing, payments, and financial management for SMBs, is acquiring Helena CRM, a conversational CRM built to run inside WhatsApp, for R$150M ($29.6M), its largest deal ever and second acquisition of 2026. The deal creates an end-to-end platform with Helena covering the sales journey from lead generation and chatbots to funnel management and Asaas handling the payments. The move pushes Asaas earlier into its customers' buying process, pairing its payment tools with autonomous AI sales agents so a small business can run everything from lead to invoice in one place, with WhatsApp the default sales channel for Brazilian businesses.
OpenAI will acquire Ona, a German startup that runs secure cloud environments for AI coding agents to execute long, multi-step tasks autonomously, for an undisclosed amount. The deal will enable OpenAI Codex agents to keep working on long jobs inside a customer's own cloud rather than being tied to a single device or session, with customers having control over access, credentials, and logging, while OpenAI supplies the intelligence and orchestration. Once the deal clears regulatory approval, Ona's team will join OpenAI and fold into the Codex group, with its technology built directly into Codex.
Blnk, an Egyptian BNPL lender that aims to serve customers traditional banks overlook, raised $37M in a funding round split between $12.5M in equity, led by Algebra Ventures, and $24.6M in debt from local banks. The company, which just turned profitable, approves financing right at the checkout counter using AI underwriting via a network of more than 3,000 merchants and over a million borrowers, most of whom had little access to formal credit before. The funding will go toward launching a credit card that lets those customers borrow beyond Blnk's own network for the first time, and a potential expansion into other countries.
Amazon raised roughly $27.5B in about 48 hours to feed its AI buildout, pairing a C$14B (about $10B) Canadian-dollar bond with a separate $17.5B Citigroup-led bank loan signed the same day. The money, which adds to the $140B cash and investments the company has on hand, will go toward Amazon's $200B planned capex earmarked for AI chips and data center development this year. The transaction was the largest-ever in Canadian dollars, breaking a record set by Alphabet a month ago when it raised C$8.5B from a four-part bond sale. Amazon has now borrowed more than $82B since the start of 2025, including debt sales in euros and Swiss francs.
Anthropic secured a $35B private credit deal led by Apollo and Blackstone to finance its purchase of Google-developed AI chips in a deal internally known as Project Big Sky. The financing runs through a special-purpose vehicle backed by chip-lease agreements, with Broadcom, the manufacturer of the Google TPUs, backstopping the senior tranches to cut borrowing costs. Google sits on nearly every side of the deal, developing the chips, guaranteeing the leases, and owning about 14% of Anthropic itself. Do these companies not have enough money as it is? Good lord…
Raptive, a creator-monetization and media company focused on premium food, family, and lifestyle publishers, acquired AlchemyAI, a stealth agentic food-intelligence startup, for an undisclosed amount. Alongside the acquisition, the company is launching Raptive Intelligence, a new unit led by AlchemyAI founder John Roa that will license Raptive's creator expertise and consumer-intent data to AI platforms, CPGs, grocers, and retailers that need authoritative content to ground AI answers and power agentic product discovery. AlchemyAI will supply the foundation, a food knowledge graph of recipes and ingredients, while Raptive provides the reach to help AI assistants plan meals, recommend products, and build shopping baskets.
Meesho, a Bengaluru-based e-commerce marketplace, is acquiring Kirana Club, a Singapore-based startup that builds software for grocery stores, and Retail Pulse Labs, its Indian subsidiary that runs a B2B marketplace linking small retailers with consumer-goods brands and distributors, for ₹202.08 crore ($21.2M). The deal deepens Meesho's reach among India's small neighborhood grocery retailers, known as kirana, and pushes the company further into merchant tooling beyond its consumer marketplace. The acquisition follows Meesho's recent public listing and intensifying competition from B2B players like Udaan and Jumbotail.
Minerva, a Brooklyn-based AI platform that helps consumer brands unify their first-party customer data and automate marketing workflows with AI agents, announced its public launch alongside a $20M funding round from investors including The General Partnership, 8VC, Lingotto Innovation, and others. The platform unifies a brand's fragmented first-party customer data, enriches it through a proprietary identity graph, and deploys agents that handle campaign creation, analysis, optimization, and reporting, with tools like an “Agentic Data Scientist,” built in collaboration with OpenAI, that lets marketers build predictive models from plain-language prompts. Minerva has signed more than three dozen customers including the NBA, and will use the funding to expand its team, build a self-service version, and move beyond its initial focus on sports, hospitality, and financial services.
Unifonic, a Saudi-based AI customer-engagement platform focused on emerging markets, acquired Segmentify, a Turkish AI personalization and behavioral-intelligence company that serves enterprise retailers and e-commerce brands across EMEA, for an undisclosed amount. The deal pairs Unifonic's conversational engagement and orchestration with Segmentify's personalization and recommendation engine to power what the companies call “agentic marketing,” which are autonomous AI agents that decide the next-best message, channel, and timing for each customer in real time rather than running manual campaigns. The acquisition also extends Unifonic's reach into the UK, Türkiye, and Germany while bringing Segmentify's personalization into the MENA and GCC markets where Unifonic is strongest.
Forage, a payments-infrastructure company that lets retailers accept government benefits like SNAP and WIC, raised $40M in a Series B round led by Mouro Capital. The company's technology now runs at more than 100,000 stores across all 50 states, including partners like Dollar General, Gopuff, DoorDash, and Uber Eats. The funding will go toward scaling its backend and growing its consumer app, which helps low-income shoppers check EBT balances in real time, find SNAP-accepting stores, and see which items qualify.
Bending Spoons, the Italian company behind digital brands including AOL, Vimeo, WeTransfer, and Eventbrite, filed for a Nasdaq IPO targeting a valuation of at least $20B. The company runs an acquire-and-revitalize model, buying up digital businesses and slashing costs, oftentimes through layoffs. It plans to use proceeds from its IPO to keep expanding, with CEO Luca Ferrari saying it has identified more than 1,000 potential acquisition targets. Bending Spoons reached profitability in its most recent quarter, reporting $27.5M in net income on $601M in revenue, up from a loss a year earlier, with much of that revenue coming from recurring subscriptions. The company was valued around $11B in an October funding round, and is aiming to launch the offering by the end of June with Goldman Sachs, J.P. Morgan, and Allen & Co leading.
Tokenz, a Tokyo-based platform that provides merchant-of-record services for digital content businesses, raised ¥1B ($6.3M) in a Series A round led by Headline Asia. Its service handles payments, tax, billing, and customer support on behalf of content sellers, taking on the tax and compliance burden so providers don't have to build global payment and international tax handling themselves, targeting gaming companies, video streaming services, and app makers across more than 200 payment methods and 224 countries and regions. The funding will go toward expanding payment options, adding AI-based fraud detection and tax features, and growing its support presence across Japan, Singapore, the US, Lithuania, and Taiwan, as a new Japanese law forcing Apple and Google to allow out-of-app payments pushes content makers toward third-party payment solutions to dodge platform fees as high as 30%.
Satispay, an Italian payments and financial-services app, is asking shareholders to approve a capital increase of up to €120M ($139.2M), with about half the raise already committed by existing backers including Greyhound, Addition, and Lightrock. The deal, which would keep the founders in control and holds Satispay's valuation above €1B, would help fund the company's expansion from payments into a full financial platform that spans investing, BNPL, pensions, and stock and ETF trading, as well as fund future acquisitions. The raise follows annualized revenue topping €116M, up 80% YoY over its last two quarters, with 6.5M users and 450,000 merchants now on the platform.
Levata, an enterprise technology provider that distributes and deploys data-capture, mobility, and point-of-sale hardware, acquired POSDATA Group, a US distributor and deployment partner for enterprise payment terminals, for an undisclosed amount. The deal moves Levata into payment hardware, adding POSDATA's terminal brands like Verifone, Ingenico, and PAX, plus its certified encryption key injection facility, to the equipment it already handles for enterprise retailers. Levata has made more than 20 acquisitions and deployed over $400M since 2014, backed by PE firm Odyssey Investment Partners.
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 happens after Trinity dies in The Matrix? Keanu Grieves