On August 4, Zillow Group cut more than 500 employees, about 7% of its global workforce. It was the company’s second round this year and its largest. CEO Jeremy Wacksman posted the news to the company blog, framing it as 「ensuring we have a disciplined cost structure and getting more efficient, with the right people in the right positions.」 Continuing to grow at scale, he wrote, requires working differently than the company does today.
Asked whether AI drove the cuts, Zillow told GeekWire no. A spokesperson said the changes were 「about better positioning Zillow for the path ahead.」
Three months earlier, the same CEO told investors Zillow was rapidly becoming an AI-native company.
This is not a shrinking company
Zillow reported Q2 the next afternoon and the numbers were strong.
Revenue of $772 million, up 18% year over year, above the high end of its own outlook. For Sale revenue rose 14% to $549M, Residential 7% to $465M, Rentals 31% to $209M, Mortgages 75% to $84M. Net loss was $4M, a 1% loss margin and 80 basis points better than a year ago. Adjusted EBITDA landed at $176M, a 23% margin, also above the top of guidance.
The company also priced the layoff: the 7% reduction is expected to produce $75M in annualized EBITDA savings.
$75M across roughly 500 seats is about $150K per seat. That sits close to the $179K figure backed out of Scripps’ cuts the same week, at two companies with nothing in common commercially. Both were removing the same salary band: mid-to-senior white-collar execution, not entry level. The Washington state detail points the same way, with 91 jobs cut in the state and senior roles taking the heaviest share.
One number is easy to skip past. Zillow had 7,058 employees as of March 31, down 10 from the end of 2025. The roughly 200 cuts it made in January were characterized as performance-related and part of the annual review cycle, and by the end of Q1 those seats had largely been backfilled. August 4 is the first time this year the company actually pushed headcount down.
Two registers, depending on the room
Lay Zillow’s own public statements out along the calendar and the split is clean.
In April, Wacksman told real estate executives at the T3 Leadership Summit that Zillow staff were being retrained to use AI in their work, and that the gains 「are small, but they’re compounding.」 In May, he told investors the company was rapidly becoming AI-native. In August, cutting 500 people, the line for the press was that AI was not the reason.
Zillow is the third instance this quarter. Patreon cut 20% in July and denied AI. Intuit cut 3,000 in May and said it had nothing to do with AI. Three companies, three denials, one structure: AI is the growth story for the capital markets and the retraining program for staff, and it is never the reason on the layoff notice.
Fortune named the pattern on August 8. Researchers call it AI washing when a company attributes cuts to AI to look forward-thinking, and the reverse, avoiding the word to dodge public backlash, is the second half of the same problem. Both are running in the same quarter, often inside different rooms of the same company.
Scripps put AI at the front of its restructuring and the stock rose 20% that day. Zillow kept AI out of its layoff and posted 18% growth the same week. Neither company said anything false. Neither account, on its own or together, adds up to a complete picture.
Why this breaks the measurement
We reported in the August WARN piece that New York State added an AI-attribution field to its WARN system a year ago, and that across more than 160 filings not one employer has used it.
Zillow shows the problem is not only the form. A company can put 「AI-native」 in its investor materials and leave AI out of its layoff statement without misrepresenting either, because the proximate trigger genuinely may be cost discipline, spending cadence, or legal expense, with AI only making those decisions easier to reach. The attribution chain snaps in the middle, and the statistics can only record the part before the break.
Every displacement index built on self-reported AI attribution is therefore biased low. Worth holding in mind whenever Challenger, WARN, or company announcements are the input.
Who is exposed
Zillow did not name the affected teams. Three things are known: 7% globally, 91 in Washington state, senior roles over-weighted.
Read against the business, the shape is inferable. The three fastest-growing lines (Rentals, Mortgages, For Sale) are unlikely to be the main source, because the company has been spending into them. The pressure sits in the supporting layer: middle-tier product and program management, the parts of internal tooling and data teams that an AI workflow consolidates, content and operations roles, and any function where a year of AI retraining now lets one person cover what two covered before. Wacksman’s April phrase, small but compounding, describes the input. August is when compounding gets collected.
For an individual the read is more direct than it is for the sector. If your employer is telling investors an AI-native story while asking your function to raise output using AI, those are two views of one decision. It reaches the earnings call before it reaches your inbox. At Zillow the gap was three months.
The rest of the category is contracting in parallel. CoStar has cut the Homes.com inside-sales team by nearly 40%. Better founder Vishal Garg stepped down as CEO on August 3. Rocket cut about 2% of the combined workforce after closing its Redfin acquisition. Listing portals and property tech are thinning their white-collar layers at the same time, and almost none of them are writing AI on the notice.
Sources
- Zillow cuts more than 500 jobs in its largest layoff of the year (GeekWire, August 4, 2026)
- Zillow Group Reports Second-Quarter 2026 Financial Results (Zillow Group investor relations)
- Zillow cuts 500+ roles ahead of Q2 earnings release (Real Estate News, August 4, 2026)
- AI Is changing work faster than the data can keep up (Fortune, August 8, 2026)
