In April 2025, GamesBeat stopped being part of VentureBeat. It spun out as a separate independent entity with Gina Joseph, previously VentureBeat’s chief strategy officer, as chief executive, and Dean Takahashi as editorial director. It now lives at gamesbeat.com under its own ownership.
That single change invalidates a large share of the pitching advice still circulating about VentureBeat, and it is the kind of thing nobody updates. If your media list has VentureBeat down as a games outlet, every gaming pitch you send there is going to the wrong company. The outlet that remains is narrower, more specific and considerably easier to get into than its reputation suggests, provided you pitch the thing it actually publishes.
Gaming Pitches Go Somewhere Else Now
What VentureBeat covers today, per its own about page, is artificial intelligence and machine learning, data science and analytics, enterprise IT and cloud computing, cybersecurity, and emerging technologies. As of 2025 it described its focus as the enterprise shift to AI, data infrastructure and intelligent security.
Read that as a narrowing rather than a broad technology remit. This is an enterprise technology publication now. Consumer apps, hardware reviews, creator economy stories and anything games-adjacent are outside scope, and the fastest way to mark yourself as someone who has not read the publication is to pitch one of them.
The company itself remains independent and founder-connected. VentureBeat was founded in 2006 by Matt Marshall, formerly of the San Jose Mercury News, raised venture funding in 2014 and 2015 from Crosslink Capital, Walden Venture Capital, Rally Ventures, Formation 8 and Lightbank, and I found no acquisition or change of control. Marshall was still authoring posts in 2026. His formal current title is not stated on any VentureBeat page, so if you are writing a pitch, address an editor rather than guessing at his role.
There Is a Form, and It Is the Only Door
VentureBeat’s guest contribution route is published, open and run through a single Typeform linked from venturebeat.com/guest-posts. That form’s own title names it as a submission to DataDecisionMakers and VentureBeat.
This matters more than it sounds. Most publications at this level either accept no outside contributions or route them through a relationship you do not have. VentureBeat publishes the door. Use it, and do not try to shortcut around it by emailing an editor directly with a finished guest piece, because the form exists precisely to manage that volume and going around it reads as not having read the instructions.
The stated topics the program wants are specific: artificial intelligence and machine learning, data infrastructure and enterprise analytics, and security. Three lanes. If your piece is not clearly in one of them, the lane problem is your problem, not the editor’s.
Submit as an editable Google Doc. The form says Google Docs only, and it means editable rather than view-only, which is a small permissions detail that silently kills submissions. The featured image needs a 2:1 ratio at a minimum of 1,200 by 600 pixels, and any images or data you include have to be properly licensed. The author bio must carry a first and last name, because bylines go to individuals rather than companies.
Write 800 to 1,200 Words, Not 2,000
The published guidance is unusually direct: “Typically, articles with 800-1,200 words tend to do well on VentureBeat.”
That is a short piece by thought leadership standards, and it is short on purpose. An enterprise technology reader scanning on a Tuesday morning wants one argument supported by something concrete, not a survey of a field. Twelve hundred words buys you a claim, a mechanism, two pieces of evidence, an honest limitation and a recommendation. It does not buy you a history of the industry or three paragraphs establishing why the topic matters.
The best structural advice I can give for this outlet is to write about a failure mode. Enterprise technology readers have heard every success story, and the pieces that land in this category are the ones explaining why a widely adopted approach breaks under a specific condition. If you have run agent infrastructure in production and know the failure that nobody writes about, that is the piece. If you want to explain why AI matters to enterprises, that piece already exists several hundred times.
VentureBeat also commits to something almost nobody else does: “You can expect to hear from us within fourteen business days.” That is a real service level on a submission form, and you should treat it as one. Diary it. Day fifteen is a no, and you are free to place the piece elsewhere without burning a bridge.
The AI Rule Is Absolute
One line on the guest post page deserves its own section because the consequences of getting it wrong are permanent: “We will not accept pieces that are flagged as written by AI.”
Note the wording. Not written by AI, flagged as written by AI. That is a detection-based rule, which means the burden sits on how your draft reads rather than on what you actually did. A piece assembled from a model’s first draft and lightly edited will read as one, and the things that flag it are the things that make writing generic in the first place: even paragraph lengths, hedged claims, transitional phrases doing no work, and no specific numbers or named examples anywhere.
The defence is not a detector-evasion trick. It is specificity. Name the system, the version, the scale, the date and the number. Describe the thing that surprised you. Include the detail that only someone who did the work would know. A piece carrying four concrete specifics per thousand words does not read like generated text, because generated text does not have access to your production incident.
The submission must also be original and unpublished, so a post already on your company blog or Medium is not eligible, and the program states plainly that “DataDecisionMakers is not the right place for marketing or promotional pieces.” The no-vested-interest rule is enforced through content rather than disclosure: if the argument only works if readers buy what you sell, it fails.
What Is DataDecisionMakers Actually Publishing?
Worth answering with evidence rather than assumption, because the program’s subject matter has drifted toward a specific set of problems and matching it raises your odds considerably.
The hub at venturebeat.com/category/datadecisionmakers runs at a steady cadence, with posts appearing through 2026 and the most recent I found dated 4 October 2026. Bylines are individual outside experts, not companies. The subject matter concentrates heavily on enterprise AI agents, retrieval-augmented generation reliability, agent identity and security, and data pipeline quality.
That is a precise content signal and you should use it. The program is not looking for AI strategy commentary. It is looking for practitioners explaining why agents fail in production, how retrieval systems degrade, what breaks when you give an autonomous system credentials, and how data quality problems surface downstream. If you have hands-on experience with any of those, you are the profile the program is built for.
To answer the question people ask about this program: nothing happened to DataDecisionMakers. It was not shut down, it did not go dormant, and it remains the primary vehicle for VentureBeat guest contributions. Posts dated within a day of writing confirm it.
Who Runs the Desk
The named editorial leadership, with titles I could verify first-party, is short.
Michael Nuñez is editorial director, a title stated on his own VentureBeat author page, and he has held the role since 2023. Carl Franzen is executive editor and continues to drive news decision-making. Karyne Levy joined as managing editor on 3 November 2025, previously at Protocol. The contact page lists reporters by beat: Franzen, Nuñez and Emilia David on AI, and Shubham Sharma on data technology.
One correction worth making because it appears in a lot of pitch guides: there is no editor-in-chief at VentureBeat. The top editorial title is editorial director. Addressing a pitch to the editor-in-chief signals a template.
The office address is 500 Sansome Street, Suite 601, San Francisco, and the contact page notes that a number of staff work remotely, which is a useful reminder that nothing is gained by trying to reach anyone in person.
Before you write, there is one piece of housekeeping that costs nothing and saves a wasted submission: confirm the piece has not already been written in your lane three times this quarter.
Open the DataDecisionMakers hub and read the last twenty posts. Not the headlines, the arguments. The program publishes at a steady enough cadence that twenty posts covers roughly two months, and by the end of that reading you will know whether your angle is fresh, adjacent to something already run, or a restatement. Editors reject far more submissions for redundancy than for quality, and redundancy is the one rejection reason you can eliminate yourself in forty minutes.
While you are there, note the shape of the pieces that ran. Observe how quickly they get to the specific problem, how early the first concrete number appears, and how they handle the author’s commercial interest, which is usually by naming it once and then never mentioning the company again. Those patterns are the house style, and matching them is a larger factor in acceptance than the topic you choose.
Separate the Editorial Door From the Sales Door
VentureBeat publishes three addresses and they do three different things. Conflating them is the single most common own-goal in technology PR.
tips@venturebeat.com takes news pitches, product announcements, funding news and corrections. This is where a staff reporter might pick up your story, and it is the right door for anything that is actually news. events@venturebeat.com handles event questions, including the Agentic Infrastructure Stack Summit scheduled for 29 October 2026 at Terra Gallery in San Francisco, and VB Transform, whose 2026 edition ran in Menlo Park in July. sales@venturebeat.com handles sponsored posts and partnerships.
That last one is paid placement. It is a legitimate product and plenty of companies buy it, but it is not editorial and it will not be presented as editorial. Decide which you want before you write to anyone, because sending a guest post draft to sales gets it priced, and sending a sponsorship enquiry to tips gets it deleted.
The practical sequence for most companies is straightforward. Funding and product news goes to tips as news, with no expectation. A practitioner argument goes through the guest post form, where you have a documented route and a fourteen-day answer. Those are separate efforts with separate content, and trying to make one piece serve both is why so many submissions read as promotional.
So which is it for you: do you have news, or do you have something you learned the hard way that nobody in enterprise AI has written down yet?