On Oct. 1, HubSpot expanded a pricing pilot it calls flexible seats-and-credits pricing to new customers across Europe, the Middle East and Africa. An earlier version ran with new customers in the Nordics and Benelux. HubSpot’s CFO told analysts in September that while it was still very early, those customers had started with smaller deals but were showing a stronger growth trajectory, which gave the company confidence to expand. When one analyst asked whether the model would go global, HubSpot didn’t address that part of the question. Existing accounts stay on their current pricing for now.
The structure has three parts. You pick one edition (Starter, Professional or Enterprise), and that single tier governs everyone in the portal. You pay per seat. GTM Seats are for the people doing marketing, sales and service work, and lower-cost Ops Seats, starting at €50 a month, are for the operations staff who maintain the system. Then usage runs on HubSpot Credits, with 5k, 10k or 15k included each month depending on edition. They reset monthly with no rollover.
There’s a lot to like here. Marketing contacts are gone as a separate charge and are now just one more record type under the account’s storage cap. Features like multi-touch revenue attribution and custom objects now sit in Professional. And buying one platform instead of assembling Hubs is an easier conversation to have with a CFO.
The tradeoff is what credits now cover. Apart from one-to-one emails and anything covered by a capacity pack, every email HubSpot sends costs 15 credits and a standard workflow action costs 3, at least for the first 10k of each. At that rate, a single newsletter to 2k contacts would use 30k credits, three times what Professional includes each month. Those flat-fee capacity packs help with budgeting, but far more day-to-day activity now draws down credits.
What stands out is how directly this ties cost to build quality. A workflow with redundant steps, or one that re-enrolls contacts it shouldn’t, used to be a housekeeping problem. Under this model it carries a price. We’ll be watching HubSpot’s next earnings call for any word on a U.S. rollout. In the meantime, it seems like a good moment to look closely at what your portal is actually running.
-Sage Levene, VP of Marketing, Hypha HubSpot Development
Open Mic
Customizing HubSpot CMS for a Bespoke Client Experience
By Amanda McIntosh, Senior Web Designer
One of the best things about developing in HubSpot is how much access we have to tweak the experience for the CMS user. A premium editing experience is a big competitive advantage top-shelf development agencies can give their clients.
If you’re like us, you probably take advantage of HubSpot’sboilerplate CMS themes as a starting point when building a custom theme. If you don’t, I’ll take a moment to sell you on it. These base themes can easily be cloned in a client’s account, or fetched locally from the @hubspot folder, and give you a great starting point for a robust custom experience. Although HubSpot has moved these themes to “Maintenance Mode” to try and shift focus to their React-based Elevate theme, base themes like Growth remain essential tools.
These themes have a robust set of base modules and templates, and can be very flexible. Here are some of my favorite ways to add another level of polish and client branding to a custom HubSpot theme:
Make a custom theme cover: You can define your theme’s cover within the theme.json file. HubSpot doesn’t have recommended dimensions for these, as the aspect ratio depends on the screen size. I’ve had the best luck using a page mockup with some elements enlarged/simplified so as to not be too busy, sometimes with a header at the top.
Remove explicit page template preview declarations unless you are making custom ones so that HubSpot can auto-generate previews.
Switch out the default module icons to match either the purpose or the client brand. Because you can use any SVG for a module icon, you have a great deal of flexibility for what you can use. Bonus points—selecting correct module categories for easy navigation.
If you use a HubSpot Boilerplate theme, replace all references to “Boilerplate” with the client name, and make sure that any placeholder content is related to your client.
Take advantage of custom sections to pre-configure modules with useful client content. Configuring the same module with different settings and content options can be a great way to speed up your client’s workflow. If you are developing locally, this can also be a great way to utilize Claude to avoid busy work—simply feed in some content and reference the correct modules and Claude can easily slot the content into the correct areas.
In addition to that list, I wanted to take a moment to touch on one of my other favorite ways to provide a premium, custom experience for clients —ensuring that your client’s brand settings are configured well.
Although this isn’t strictly a “developer” task, approaching this task with a design-focused mindset allows you to make sure your client will have everything they might need to make quality branded content. If you can, make sure you include light and dark vector logos (.SVG) for web, as well as small PNG logos for use in email. Ensure the your brand colors not only include the client’s primary colors, but also branded very light and very dark colors that can be used in lieu of HubSpot’s default light grey and dark blue text.
I hope these give you some good ideas for how to provide your clients an even better CMS experience. Want to partner with a team that can bring this level of attention to detail to your HubSpot theme? Reach out! We’d love to work with you.
Get ready, we’re about to live in vertical world full time.
“TechCrunch reports that some users in India are getting a Facebook test that, upon opening the app, drops them directly into an immersive, full-screen video feed. Meta hinted that this ‘experience’ would be coming to ‘video-heavy countries’ first, followed by a potential US expansion in 2027.
“TikTok’s — and by extension, its copycats’ — focus on video has reshaped all of the industries that depend upon such platforms for distribution. Podcasts have become videos, which have become clips. TV shows have become ‘microdramas’ and movies have been chopped up into 50 parts...It’s not clear if users actually prefer video over text or images, or if they simply engage with it because it’s what social media platforms have placed directly in front of them — engagement is engagement, so platforms likely don’t care either way.”
We’ve long known that SERPs prefer and reward pieces that have a person behind the byline (when executed properly), but this warning takes it a step further. Real people, not real fake AI people.
“Google added a new paragraph under the who, how and why section of that document with the following lines:
“‘Avoid using deceptive authorship information.’
“‘Fabricating creator profiles (such as by using AI-generated headshots, made-up names, or false credentials to make content appear as if it was written by human experts) is a form of deception.
“‘Any form of deception makes a page untrustworthy to both users and our automated quality systems, and is a signal of a low-quality page.’”
Hypha Highlights
Consider a contact who fills out a form. A rep calls, no one picks up, and the status changes to “left message, not answered.” If nothing else is built into the process, that is the end of the contact’s journey. The lead qualified well enough to get a call, but no email follows, no task comes due, and nobody has decided whether they should hear from you again next week or next quarter.
Marketing automation lead management is how you prevent that. It covers the stretch between a lead qualifying and being ready for sales, including how their treatment changes along the way and when it ends. The examples below follow an invented food business (no client is involved) that sells burgers, pizza, salads, and açai bowls, which keeps the HubSpot setup easy to follow.
“We have decided to reduce the size of our team by ~7% and will be saying goodbye to nearly 660 HubSpotters. We did not make this decision lightly and approached it with the seriousness and care it deserves.
“We are making three changes to how we operate...Organize around customer outcomes. Build a flatter organization with fewer layers. Create agile teams with clear ownership.
“This is not driven by AI-related efficiencies. We believe in a world where AI helps make us more productive and we will continue to invest to make that happen. This change is about aligning our organization with our strategy and how we need to operate going forward.
“This is not simply a cost-cutting exercise. We have been thoughtful about how we invest in AI while balancing growth and profitability. We have been disciplined about growing headcount slower than revenue, and we will continue to do so. This change is about where and how we invest so we can put more resources behind our biggest opportunities.
“None of that makes today easier. We are saying goodbye to colleagues who have helped build HubSpot. We owe them our gratitude, our respect, and our support through this transition.”
AI in Action
News, updates and tools from the AI industry.
OpenAI’s annualized revenue is about $50 billion—roughly $20 billion less than the $70 billion figure widely reported in September—a discrepancy stemming from how AI companies account for sales made through cloud partners. Anthropic records the full customer payment as revenue and lists the cloud provider’s cut as an expense, while OpenAI books only its share; the higher $70 billion figure was reportedly an attempt to “gross up” OpenAI’s number to match Anthropic’s methodology for investors wanting an apples-to-apples comparison.
Google Cloud unveiled a universal Gemini agent for the workplace that can run tasks lasting hours or days while maintaining memory and identity across apps—configured as either a personal assistant or a named digital “coworker” with its own Workspace account, email, calendar, and Drive storage. On the consumer side, a startup called Tab emerged from stealth at a $300 million valuation, joining assistants like Instinct and Meta’s Muse; users text it on iMessage or WhatsApp to handle everyday tasks, and the company pitches trust as its differentiator, keeping user data out of training and card details in a vault it never sees.
Anthropic made two moves this week. It launched Claude Haiku 5.5, the first update to its smallest, cheapest model in nearly a year, cutting per-token prices by roughly 75% on average. Separately, Anthropic published its annual Usage Policy update, consolidating rules against deceptive influence campaigns, sharpening its elections and surveillance restrictions, spelling out “human in the loop” requirements for high-risk and autonomous-hardware use, and adding a narrow prohibition on sustained, needless abuse of its models.
Crunchbase data shows that some of the AI industry’s fastest-growing startups are becoming serial acquirers, buying smaller companies to fill product gaps, enter new markets, and bring specialized teams in-house—pushing acquisitions of AI startups by other venture-backed AI companies to 195 through late September, 14% more than all of 2025, even as the number of buyers grew just 2%. OpenAI leads by far with 20 such deals over three years (10 this year), followed by repeat buyers including Anthropic and legal-tech startups Legora and Harvey.
DeepSeek is close to securing at least 80 billion yuan (roughly $12 billion) in its latest funding round—blowing past its initial 50 billion yuan target amid strong investor interest—ahead of a landmark IPO expected in early 2027, with Tencent and battery giant CATL among the largest backers. The round, which could push the company’s valuation to around $75 billion, follows the successful release of its V4 Flash model that reset cost-versus-performance expectations against Anthropic and OpenAI, as founder Liang Wenfeng pledges to keep developing open AI models while pursuing artificial general intelligence.
In an interview with Politico’s Decoded, OpenAI CEO Sam Altman said there remains “a lot of daylight” between his company and rival Anthropic on AI regulation, arguing that “the world should accept some bad things happening for the benefits of this technology” and that AI should remain broadly accessible rather than tightly controlled.
Former FTC chair Lina Khan said on ABC’s “This Week” that AI companies attempting to self-regulate would be repeating the “proven failure” of Big Tech’s self-regulation in the early social-media era, calling it “an enormous mistake” to follow that playbook. Citing a string of cybersecurity incidents in which AI agents hacked other companies and governments, Khan argued it is “laughable” to suggest that adopting new technology exempts companies from existing laws, as the Trump administration continues to favor industry self-regulation over formal AI rules.
“The AGI Chronicles is the definitive, behind-the-scenes account of the trillion-dollar sprint to create AGI. Drawing on more than 150 interviews with insiders at OpenAI, Anthropic, and Google DeepMind, as well as unpublished company documents and research memos, Roose reconstructs the most consequential decade in tech history. He takes us inside the rooms where researchers watch large language models learn to talk, reason, and code at superhuman speed. He brings us to the White House, where top officials grapple with a technology that is improving too fast to understand, let alone regulate. He offers new details on the disagreements that led the founders of Anthropic to defect from OpenAI, and fresh reporting on Sam Altman’s dramatic firing—and equally dramatic reinstatement—as CEO. He gives us the inside story of Google’s frantic attempt to catch up to the AI frontier. And he ushers us into the San Francisco AI scene, where ‘accelerationists’ in hacker houses and ‘doomers’ at AI safety summits battle each other for supremacy.”
How can we help you?
Case Study: HubSpot Customer Agent
A family-run manufacturer selling direct to homeowners had outgrown their ability to field customer inquries. Customer questions came in by email, Facebook, and website forms—all landing with the same few family members also running sales, quoting, production, and shipping. Messages went unanswered, quote requests sat, and a backlog built across channels nobody had time to check.
The agent was part of the scope from the start, but it could only work in channels HubSpot could see. We built the foundation first: Help Desk, shared inbox, email and Facebook connected, forms feeding the same view, routing and follow-up mapped. Then pointed Customer Agent at both new volume and the existing backlog—customers waiting got replies instead of being written off.
2,732 conversations handled, 1,601 resolved by the agent, 59% resolution rate. One week hit 182 resolutions. Growth no longer lands entirely on the family.
Small team answering the same questions across too many channels? Contact Hypha to talk through what that system could look like.
