OpenAI Phishing Scam Uses Fake SpaceX, Anthropic Event Invites to Steal Microsoft and Google Accounts

An OpenAI phishing scam is targeting corporate Microsoft and Google accounts with fake event invitations using the names of OpenAI, SpaceX, and Anthropic. Security researchers at Push Security documented the campaign on November 6, 2026. It spreads malicious links disguised as conference registrations and exclusive webinar invitations.

The Lure: Free ChatGPT Pro, Grok Heavy, and SpaceX Shares

Push Security researchers found that the operators are building fake websites for events that don't exist. The list includes an OpenAI Academy livestream, a private SpaceX briefing on AI marketing, and an Anthropic presentation at AWS re:Invent.

The pitch is a freebie. Visitors to those pages are told they can claim a complimentary ChatGPT Pro subscription, get access to Grok Heavy, and even receive SpaceX shares. That last offer is the sharpest hook: it's tied to a company whose founder routinely attaches rockets to tweets. Here's the thing: the reward is a pretext, not the payload. To claim anything, visitors are pushed into a login flow, where the credential theft happens.

How the Login Trap Actually Works

The campaign doesn't rely on a single trick. According to the research, attackers combine device-code phishing, reverse proxies, and techniques designed to bypass multi-factor authentication (MFA). Device-code phishing is the part to understand. Rather than using a fake password box to harvest a password, the attackers give victims a code and tell them to enter it on what looks like a legitimate Microsoft or Google authentication page.

The code comes from the attacker, and completing the flow signs them into the victim's real account. The victim sees a genuine login, but the session belongs to someone else. Reverse proxies add another layer: the credential-harvesting page runs through infrastructure that mirrors a real corporate sign-in portal, making the URL harder to inspect and the certificate harder to read at a glance. Combined with MFA bypass, these tactics can leave a victim fully compromised, even as they believe they've simply signed in.

Why AI Brands Are the Bait Now

The timing isn't accidental. Employee interest in AI tools makes them a high-success lure because access requests are routine. A message offering a free ChatGPT Pro seat or an Anthropic briefing looks like an opportunity, not an invoice, and can avoid the suspicion a payroll-themed phishing email might trigger.

That helps explain the targeting pattern. The victims are corporate Microsoft and Google account holders, not people using personal inboxes. A compromised work identity can expose email, cloud storage, code repositories, and downstream access to colleagues. Techradar first reported the findings, detailing Push Security's research and the credential-harvesting methods behind the fake event pages.

What Organizations Should Do Now

For security teams, the defensive lesson is specific and actionable. Block obvious lure phrases -- free ChatGPT Pro, Grok Heavy, SpaceX shares -- at the mail gateway, and treat invitations to events that don't exist as hostile by default. If a user has already entered a code on a page reached through such an invitation, revoke the session immediately. A password reset alone isn't enough, since device-code flows leave an active token behind.

For individuals, the giveaway is in the request itself. No legitimate OpenAI, SpaceX, or Anthropic promotion requires you to sign in to claim a free subscription. And no conference registration asks you to approve a device code you didn't initiate. That pattern, not polished branding, is the reliable signal.

Where This Heads Next

The campaign points to a lasting trend, not a one-off email blast. Lures based on AI brand launches, model releases, and developer briefings will keep spreading because the interest is real, and the audience is already looking for exactly the kind of opportunity being offered.

Expect copycats. The same playbook -- fake event, free reward, device-code login, reverse proxy front -- is cheap to replicate and easy to reskin around any trending announcement. Security teams should assume the campaign's specific brand names will be swapped out within weeks. That makes the device-code approval request, not the OpenAI logo, the thing people need to learn to spot. The takeaway: if an AI-branded invitation asks you to approve a sign-in code, you're looking at the OpenAI phishing scam pattern, not a giveaway. Close the tab and verify through a channel you already trust.

FAQs

What is the OpenAI phishing scam described in this report?

It's a credential-theft campaign that uses fake event pages bearing the OpenAI, SpaceX, and Anthropic names to lure employees into a fraudulent login flow. Corporate Microsoft and Google accounts are the primary targets.

Which accounts are being targeted?

According to Push Security's research, the campaign targets corporate Microsoft and Google accounts rather than personal consumer email.

What do victims receive if they click the link?

They're offered a free ChatGPT Pro subscription, access to Grok Heavy, and SpaceX shares. To claim the reward, they must sign in, which is where the theft occurs.

How do the attackers bypass multi-factor authentication?

They combine device-code phishing with reverse proxies. The victim approves what looks like a legitimate sign-in while the attackers create a session they control.

What should an employee do after clicking?

Revoke active sessions and report the message to the security team promptly. Device-code phishing can leave a valid token in place even after a password change.Techradar