Intro
OpenAI is no longer the scrappy research lab that surprised the world with ChatGPT. It is a hyper‑funded, hyper‑scrutinised platform that still has a surprisingly basic problem: what, exactly, is the business? Two recent tuck‑in acquisitions – personal finance startup Hiro and business talk‑show producer TBPN – look small on the surface, but they reveal the company’s deeper anxieties: shaky monetisation, fierce enterprise competition and a growing trust deficit.
This piece looks beyond the headlines to ask what these moves say about OpenAI’s strategy, its rivalry with Anthropic, and what all of this means for companies betting their own products on US‑built AI models.
The news in brief
According to TechCrunch’s Equity podcast, OpenAI has quietly bought two young companies.
First, it acquired Hiro, a personal finance app launched only two years ago. Hiro is shutting down its consumer product, with OpenAI mainly absorbing the team – a classic acqui‑hire. The founder has a track record in consumer apps, which observers read as OpenAI fishing for new, stickier products than a generic chatbot.
Second, OpenAI purchased TBPN, a fast‑growing business talk show and media startup. Publicly, TBPN is supposed to keep editorial independence, but the team is being placed under OpenAI’s public policy and communications structure – a detail that worries media watchers.
On the podcast, TechCrunch’s reporters frame these deals as responses to two existential challenges: building something people will reliably pay for beyond ChatGPT’s current subscriptions, and repairing or reshaping OpenAI’s public image amid rising criticism and a recent investigative piece in The New Yorker.
Why this matters
Strip away the PR and these are defensive moves.
On the product side, Hiro speaks to a simple reality: ChatGPT is wildly popular but not obviously a durable, high‑margin business. Consumer subscriptions plateau, API usage is price‑sensitive, and enterprises increasingly treat large language models as interchangeable components. Pulling in a team that has shipped consumer products in a brutally competitive space like personal finance is a bet that someone inside OpenAI can design an AI‑native product users cannot easily abandon.
Who benefits? The Hiro team gets a soft landing and access to frontier models. OpenAI gets scarce product and growth talent that has lived through the grind of acquiring and retaining paying users, not just racking up MAUs during a hype wave.
TBPN is the mirror image: less about product, more about narrative control. OpenAI faces intensifying scrutiny over safety practices, corporate governance and its cosy relationship with big platforms and governments. Owning a slick business talk show – and embedding it inside policy and comms – is a way to shape the conversation upstream, before critics and regulators do.
Here the winners are OpenAI’s communications team and TBPN’s founders, who gain financial security and distribution. The losers could be media independence and, indirectly, customers who rely on robust, critical coverage to evaluate a vendor.
Competitively, all of this creates surface area for rivals. Every hour OpenAI spends on media strategy is an hour Anthropic, Google and open‑source players can spend improving models, developer experience or pricing. These small acquisitions are symptoms of a deeper strategic tension: is OpenAI a research lab, a consumer app company, a cloud‑style platform, or a political actor? Trying to be all four is exactly what gives competitors room to specialise.
The bigger picture
Zoom out and these deals fit three broader trends in the AI industry.
1. The great verticalisation.
Frontier model providers are racing to move closer to end‑user workflows. Microsoft has Copilot across Office; Google pushes AI helpers into Workspace and Cloud; smaller players bundle models into niche tools for legal, design, and customer support. OpenAI itself has experimented with plugins, a GPT Store and bespoke enterprise offerings.
Hiro sits squarely in this pattern: a vertical use case – personal finance – where AI could, in theory, manage flows of data, nudge behaviour and lock in long‑term users. The lesson from the cloud era is clear: owning the core platform is valuable, but owning the workflow and data that sit on top can be even more defensible.
2. Models are commoditising faster than expected.
The conversation TechCrunch reports from the HumanX conference – developers casually saying that ChatGPT is fine but raving about Anthropic’s Claude Code – captures a dangerous reality for OpenAI: the wow factor of being first has faded. Engineers compare latency, reliability, coding assistance, safety constraints and total cost of ownership. If Claude is better for coding this quarter, they will switch.
That pushes OpenAI to search for differentiated, consumer‑facing products and deeper enterprise relationships. Acqui‑hiring proven product builders is a rational response, but it also signals that large‑scale research alone is no longer enough.
3. Owning the megaphone.
Tech giants have long tried to shape narratives via blogs, podcasts, house‑media and sponsored conferences. Buying TBPN is merely the AI‑era version of that instinct. What is new is the level of public suspicion: when the most powerful model provider in the world owns a talk show that covers tech and business, claims of ‘editorial independence’ naturally raise eyebrows.
We have seen similar dynamics before, from oil companies sponsoring climate content to crypto exchanges funding industry media. It rarely ends well for trust. For OpenAI, that is an existential risk: its entire value proposition is built on users trusting that these systems are safe, reliable and not subtly steering them.
The European / regional angle
For European companies and policymakers, these developments crystallise several concerns.
First, concentration of power. The EU’s Digital Markets Act and Digital Services Act were designed for the last generation of platforms; the AI Act extends that logic to general‑purpose AI models. A single US‑based vendor that not only provides core models, but also owns downstream apps (like a future finance product) and influential media channels, is exactly the scenario regulators worry about: vertical integration plus informational influence.
Second, cultural expectations around privacy and independence are different in Europe. A personal finance assistant built on US‑hosted models raises sharp questions about financial data processing, GDPR compliance and data localisation. Enterprises in banking, insurance or energy will ask very different questions than a Silicon Valley startup when choosing between OpenAI, Anthropic, Google or European contenders.
Third, media capture is a sensitive topic. Many EU states already fund public broadcasters precisely to guarantee a minimum of independent coverage. When an AI lab buys a business talk show, European regulators may not intervene directly, but it feeds into ongoing debates about media pluralism and platform power.
On the opportunity side, this is a chance for European AI companies – from open‑source model communities to domain‑specific vendors – to differentiate on governance and trust, not just raw capability. A European bank may prefer a slightly less capable model run on EU soil, with clear auditability and no entanglement with a corporate PR machine in San Francisco.
Looking ahead
What happens next will tell us a lot about OpenAI’s true priorities.
If the Hiro team quietly disappears into generic product roles, this was purely a talent grab. If, instead, we see OpenAI roll out a branded, AI‑driven personal finance or productivity assistant with deep hooks into users’ daily lives, we will know the company is serious about building vertical consumer products alongside its platform.
On TBPN, the key signal will be distance. Do episodes start featuring OpenAI executives disproportionately, framing regulatory debates in the company’s preferred language? Do critical voices about safety, labour impacts or competition quietly vanish from the guest list? The more TBPN looks like a house organ over the next 12–24 months, the more it will undermine both the show and OpenAI’s credibility.
Commercially, expect OpenAI to lean harder into enterprise and developer offerings: more aggressive courting of Fortune 500 CIOs, tighter integration with existing productivity suites and perhaps pricing moves designed to undercut Anthropic on coding and enterprise workloads. The battle for being the ‘default’ AI provider inside large organisations is only beginning.
The wild card is regulation. As the EU AI Act phases in and other jurisdictions sharpen their rules, OpenAI’s dual role as infrastructure provider and media actor will be scrutinised. A serious safety or transparency scandal, amplified by the perception that the company also controls parts of the narrative, could accelerate calls for structural separation between AI labs and media properties.
The bottom line
These acqui‑hires will not change OpenAI’s trajectory on their own, but they are revealing. A company truly confident in its business model and public legitimacy does not feel the need to buy both a personal‑finance app team and a talk‑show studio in the same news cycle. The real existential questions for OpenAI are not about model quality – they are about economics and trust.
For developers, founders and CIOs, the takeaway is simple: enjoy the capabilities, but diversify your dependencies and pay close attention to who controls the narrative around the tools you rely on.
