As of December 14, 2025, one question keeps surfacing across venture circles, secondary-market platforms, and AI industry chatter: what is Scale AI actually worth right now?
The most recent widely reported headline number still points to about $29 billion, tied to Meta’s blockbuster minority investment and expanded commercial relationship with Scale AI in mid‑2025. [1]
But in the past several days, fresh reporting and updated private-market pricing signals have pulled that figure into a much more complicated—and much wider—range.
This is not unusual for late-stage private companies, but Scale AI’s case is especially volatile because the company sits at the center of a sensitive supply chain: human-labeled data and model evaluation for frontier AI labs—many of which compete directly with Meta.
Below is the most current valuation picture, what’s driving the divergence, and what investors and industry watchers are likely to focus on next.
Scale AI valuation today: why the market is showing more than one number
There is no single “official” Scale AI valuation that updates daily the way a public stock does. Instead, today’s valuation conversation is built from three buckets:
- Last disclosed financing / transaction valuation (the “paper” valuation from Meta’s deal)
- Secondary-market and brokered pricing indicators (derived prices, limited transactions, illiquidity discounts)
- Media reports and investor mark-down narratives (often reflecting sentiment and selective datapoints)
Here’s what those buckets look like as of Dec. 14, 2025:
1) The headline benchmark: “~$29B” from the Meta deal
Scale AI announced in June 2025 that Meta’s “significant new investment” valued Scale at over $29 billion and expanded their commercial relationship. [2]
Reuters similarly reported Meta investing for a 49% stake in a deal valuing Scale AI at $29 billion (with sources citing $14.3B). [3]
2) Secondary-market signals (updated recently)
Private-market pricing services don’t always agree—because the underlying transactions are sporadic, often small, and constrained by company approval and transfer restrictions.
Still, the newest visible indicators include:
- Forge (Dec. 14, 2025): shows a Forge Price of $20.00 per share for Scale AI as of 12/14/2025, described as a derived price intended to provide valuation insights; Forge also lists post‑money valuation figures tied to the Meta round. [4]
- Nasdaq Private Market: says its Tape D® estimate put Scale AI at $11.6 per share last month, based on “market activity, publicly-sourced valuation data, and other proprietary information.” [5]
- Premier Alternatives: lists a current secondary-market price of $19.64 per share and a current valuation of $18.9B (with the note that “last funding price” is not available on the page). [6]
- PM Insights: states Scale AI shares were valued at a 61.87% discount to its Series G‑1 post‑money valuation as of Dec. 1, 2025. [7]
These numbers don’t perfectly line up—and that mismatch is the point: private-market “valuation” is often a range, not a single truth.
3) The downside narrative surfacing this week: “as low as $7.3B”
In the past few days, Business Insider reported that private valuations have dropped from $29B to as low as $7.3B, in the context of turmoil following Meta’s investment and Alexandr Wang’s move. [8]
That doesn’t mean Scale AI officially “is worth $7.3B.” It does mean that some private market participants are marking it at far lower levels than the Meta-deal headline number—particularly if they believe the company’s growth profile, customer mix, or strategic independence has weakened.
What changed in the last days: new reporting refocuses the valuation debate
The sharpest near-term catalyst is the new wave of reporting about operational strain after Meta’s deal.
Business Insider’s recent report describes pressure inside Scale AI’s gig platform operations, contractor morale issues, shifting workloads, and client pullback—while also noting Scale’s claims about profitability and moves into defense and robotics. [9]
In valuation terms, that story matters because it maps directly to the variables investors use to justify premium multiples:
- Revenue durability (are the biggest customers sticking around?)
- Margin structure (is specialized data work scaling profitably?)
- Strategic risk (is Scale now seen as “too close” to Meta?)
That last point—strategic risk—has been central ever since the Meta investment closed.
The Meta factor: a $29B valuation—and an immediate neutrality problem
Meta’s investment was designed to do two things at once: provide capital and deepen the partnership—while also bringing Scale AI’s co-founder CEO Alexandr Wang into Meta’s AI leadership orbit. Reuters reported that the deal valued Scale at $29B, involved a 49% stake, and that Scale’s strategy chief Jason Droege would serve as interim CEO. [10]
From a valuation perspective, the Meta tie-up created a paradox:
- On one hand, it’s hard to get a stronger “strategic validation” than a multi‑billion‑dollar Meta check at ~$29B.
- On the other hand, Scale AI’s business historically benefited from being a trusted, neutral supplier to multiple frontier labs.
Reuters reported that Google—Scale AI’s largest customer—planned to cut ties after news of Meta taking a 49% stake, with sources describing how labs sought “neutrality” and how the relationship could expose sensitive roadmaps. [11]
Reuters also put numbers on the customer concentration dynamic: Scale AI generated $870M in 2024 revenue, and one source said Google planned to pay about $200M in 2025 (while Google spent about $150M the year prior). [12]
This is the crux of why Scale’s valuation can look like $29B in one framing—and dramatically lower in another:
- If Scale remains the default high‑quality data and evaluation vendor across many labs → premium valuation logic holds.
- If neutrality is impaired and key labs pull work in‑house or shift to rivals → revenue quality and multiple compress.
Defense work was supposed to diversify Scale’s story—then a major contract went elsewhere
One reason investors historically assigned Scale AI a higher valuation is the idea that the company could expand beyond commercial AI labs into defense and public sector work (where data quality, security, and procurement cycles can create sticky relationships).
Scale announced a five-year, $100 million ceiling enterprise agreement with the U.S. Department of Defense’s CDAO in September 2025. [13]
However, in late November, Observer reported that Enabled Intelligence won a seven-year contract with a ceiling of $708 million supporting Maven and related missions—beating larger rivals including Scale AI. Observer notes the award became public after the end of a government shutdown, and that Bloomberg first reported it. [14]
Scale’s statement to Observer emphasized that its “public sector momentum is accelerating independent of a single award,” pointing to other Pentagon and Army contracts. [15]
Why this matters for valuation:
- A $708M ceiling contract is not guaranteed revenue, but it is a powerful signal about who is winning high‑stakes labeling work.
- Losing a marquee award can influence how investors perceive Scale’s moat—especially if the company’s diversification narrative (away from hyperscaler AI labs) is central to the bull case.
Why private-market valuation estimates diverge so sharply
Readers often assume secondary markets should “tell the truth” about valuation. In private companies, they often tell multiple partial truths at once.
Here are the biggest mechanics behind the spread you’re seeing for Scale AI right now:
Different share classes can imply different effective prices
Preferred shares in late-stage rounds can carry terms (liquidation preferences, protections, conversion rights) that are not comparable to common shares sold by employees. A “$X per share” print is not automatically equivalent across share types.
Illiquidity discounts are real—and can deepen quickly
Even if a company is “worth $29B” on paper, actual buyers in a thin market may demand significant discounts to compensate for:
- transfer restrictions,
- limited disclosure,
- long time-to-liquidity,
- and the risk that the next financing reprices the company.
That logic is consistent with PM Insights describing a large discount to the Series G‑1 post‑money valuation as of early December. [16]
Pricing vendors use different methodologies
Nasdaq Private Market’s Tape D® is explicitly described as an estimate based on multiple inputs including proprietary information. [17]
Forge’s “Forge Price” is also positioned as a derived price “to provide insights into current valuation.” [18]
These are useful indicators, but they are not the same as a single, centralized exchange price.
Scale AI valuation timeline: how it got here (and why the spread is unusual)
A simplified valuation arc (based on reported rounds and deals) looks like this:
- May 2024: Reuters reported Scale AI raised $1B led by Accel, valuing it at nearly $14B, with participation from Nvidia, Amazon, and Meta. [19]
- March 2025: Reuters reported Scale AI was seeking a valuation as high as $25B in a potential tender offer (per a Business Insider report cited by Reuters). [20]
- June 2025: Meta’s investment valued Scale AI at ~$29B, with Meta taking a 49% stake and Wang moving to Meta. [21]
- Dec 2025: Secondary-market pricing services and media narratives now imply a broad range—from near the last-round headline to substantially lower marks, including reports “as low as $7.3B.” [22]
The standout feature in late 2025 is not just the number—it’s the dispersion. That dispersion usually signals a market trying to answer a binary question: Is Scale AI still a neutral infrastructure layer—or has it become strategically tethered to Meta in a way that shrinks its commercial surface area?
Outlook and forecast: what could move Scale AI’s valuation next
No one outside the company can publish a definitive “target price,” but the drivers that will likely determine whether Scale AI trends back toward the $29B headline—or stays priced at a discount—are becoming clearer.
1) Customer retention and concentration
The market will watch whether Scale can keep (or replace) major commercial customers amid neutrality concerns raised in reporting about Google and other labs stepping back. [23]
2) Revenue quality vs. volume
Scale’s work increasingly involves specialized, high-skill annotation and evaluation (often expensive per task). That can support margins—but only if demand stays strong and quality remains defensible.
3) Defense and government execution
Winning and delivering public-sector work is a credibility anchor. But the Enabled Intelligence win over Scale for a major ceiling contract underscores that defense is competitive—and that procurement outcomes can reshape the narrative quickly. [24]
4) Broader AI valuation mood going into 2026
Even as private AI valuations soar across the market, caution is rising. Investopedia recently highlighted Bill Gates warning that not all AI companies will succeed and that some valuations may not be sustainable. [25]
At the same time, Reuters reported investor interest in other AI verticals (like robotics) supporting big valuation rounds—evidence that capital is still available for the right story. [26]
5) Liquidity path: tender offers, secondaries, or IPO?
Multiple platforms note Scale AI is still private, and there is no announced IPO date. [27]
Absent an IPO, valuation debates will continue to be mediated through tender offers, negotiated secondary transactions, and pricing estimates that can swing with sentiment.
Bottom line: Scale AI valuation on Dec. 14, 2025 is best described as a range, not a point
If you need a single sentence to summarize Scale AI’s valuation “today,” it’s this:
Scale AI’s last major deal still anchors around ~$29B, but late‑2025 private-market signals and recent reporting show a much wider valuation band—reflecting uncertainty about customer neutrality, competitive pressure, and the durability of Scale’s post‑Meta growth story. [28]
For Google News and Discover readers, this is the key takeaway: Scale AI’s valuation debate is no longer just about AI hype—it’s about trust, access, and who controls the data supply chain behind the models.
References
1. scale.com, 2. scale.com, 3. www.reuters.com, 4. forgeglobal.com, 5. www.nasdaqprivatemarket.com, 6. www.premieralts.com, 7. www.pminsights.com, 8. www.businessinsider.com, 9. www.businessinsider.com, 10. www.reuters.com, 11. www.reuters.com, 12. www.reuters.com, 13. scale.com, 14. observer.com, 15. observer.com, 16. www.pminsights.com, 17. www.nasdaqprivatemarket.com, 18. forgeglobal.com, 19. www.reuters.com, 20. www.reuters.com, 21. www.reuters.com, 22. www.businessinsider.com, 23. www.reuters.com, 24. observer.com, 25. www.investopedia.com, 26. www.reuters.com, 27. www.premieralts.com, 28. www.reuters.com
