AI Says These Three AI Startups Are Undervalued

Using AI to evaluate other AI companies feels like stepping into uncharted territory—equal parts fascinating and unsettling. Welcome to 2025, where the future evaluates itself.

We've encountered countless AI companies lately, so I decided to explore three early-stage startups that our AI (which we use for target identification in our VC funds) flagged as both disruptive and significantly undervalued. Since these companies fall outside our current investment themes, we have no financial stake in their success. I filtered for AI startups based in the US, last valued under $50M in 2024, that our AI identified as trading well below their intrinsic value.

Pump: The "Costco of Cloud Computing"

Last valued: $40M in May 2024 (Pitchbook)

Our AI's intrinsic valuation estimate: >$100M

Website: pump.co

Pump is a San Francisco-based cloud cost optimization platform founded in 2022 that uses AI and group buying to help small and medium-sized businesses slash their cloud computing expenses by up to 60%.[1] They aggregate demand to unlock enterprise-level discounts that individual smaller companies couldn't access alone.

The cloud computing market favors big enterprise clients through pricing models that leave startups and small businesses stuck paying premium prices or committing to unpredictable long-term contracts. It's like being forced to buy a family-size Costco membership when you only need a single avocado—exactly the kind of market inefficiency that creates disruption opportunities.[2]

What makes Pump compelling isn't technical superiority—they're not building better servers or databases. Instead, they've developed a model letting small companies band together for big company discounts without long-term commitments. They've found a way to pool demand from customers who were either priced out or poorly served by current offerings.[3,4,5]

With hundreds of customers already signed (including 50+ YC companies),[6] Pump's approach merits attention as it evolves in a market where big providers have major advantages but also significant blind spots.

Our AI values this business model innovation so highly that it estimates Pump's true worth at more than 2.5x its last funding round—the largest discrepancy among our three highlighted companies.

Acclinate: Hopping Past Pharma's Blind Spots

Last valued: $40M in May 2024 (Pitchbook)

Our AI's intrinsic valuation estimate: >$90M

Website: acclinate.com

Acclinate is a Birmingham, Alabama-based digital health company founded in 2020 that increases diversity in clinical trials by connecting pharmaceutical companies with communities of color through AI-powered predictive analytics and authentic community engagement platforms.

The pharmaceutical industry faces a critical data quality problem: clinical trials systematically exclude key populations and variables, undermining the reliability of medical research. For example, one study found that less than 7% of clinical trials account for hormonal fluctuations in study design or analysis despite hormones significantly affecting drug metabolism, and only 28% of reproductive health technology companies reportedly base their research on diverse populations.[7] This means that treatments developed and approved through clinical trials may not work effectively—or may even be harmful—for large segments of the population who weren't adequately represented in the original research. The problem has both ethical and practical dimensions: medical treatments often work less effectively for some people just because clinical trials didn't include similar populations.[8]

What makes Acclinate notable is how they combine technology with grassroots community engagement. While many health tech companies believe an app alone can fix systemic inequality, Acclinate pairs AI analytics with real-world community-building.[9, 10] This matters because healthcare solutions often fail without cultural context and trust. Our AI values Acclinate at more than double its last funding round.

GPTZero: Whooo Wrote This Content?

Last valued: $40M in June 2024 (Pitchbook)

Our AI's intrinsic valuation estimate: >$60M

Website: gptzero.me

GPTZero is a New York-based AI detection platform founded in 2023 that identifies whether content was written by AI or humans. Initially focused on education to help teachers identify AI-generated student work, they've expanded to other markets.

GPTZero represents a "second-order disruptor"—existing only because of other disruptive technologies like ChatGPT and Claude. It's the technological equivalent of opening an umbrella shop because someone else invented rain. What makes it interesting is representing a new market category impossible before LLMs became widely available.

As AI-generated content proliferates across education, publishing, and hiring, GPTZero establishes trust through detection capabilities. Founded by Edward Tian at Princeton, it has quickly become a leader in AI detection.[13] There's beautiful symmetry in using AI to detect AI—like fighting fire with fire, I suppose.

Our AI values this company more than 50% above its last funding round—the most conservative increase among our three startups, but still significant. The premium reflects both potential upside and risk that LLM providers might eventually integrate detection capabilities themselves.

According to TechCrunch, GPTZero reports impressive growth—users increased from 1 million to 4 million in a year, with revenue growing 500% in six months.[14] The company claims profitability and recently secured a $10 million Series A led by Footwork. They've expanded beyond schools to government procurement agencies, grant-writing organizations, and companies labeling training data for AI systems.[15]

GPTZero's approach follows classic disruption theory—start with a focused market and gradually expand as technology improves.[16] That said, GPTZero isn't alone in this space. Competitors like Winston AI, Originality.ai, and Copyleaks are vying for market share. There's also real possibility that big LLMs like OpenAI, Anthropic, or Gemini might build AI detection directly into their models, potentially making standalone detectors obsolete.

The open question is whether GPTZero will continue growing or see its functionality absorbed by LLMs themselves. Meanwhile, GPTZero is a great example of a fast disruption, built upon another fast disruption... and the shockwaves continue.

Why These Companies Have Disruptive Potential

From a disruption theory perspective, these three companies share several characteristics that make them worth watching:

They're targeting overlooked segments or creating new markets. Pump serves smaller companies big cloud providers aren't focused on; Acclinate engages communities excluded from medical research; GPTZero addresses trust needs in AI-generated content that didn't exist until recently. This pattern of finding unserved customers is central to disruption theory, as Christensen repeatedly emphasized.[21]

They create difficult choices for existing incumbents. Each startup puts established players in positions where competing effectively would mean hurting existing business models.[22] Cloud providers would need to rethink pricing; pharmaceutical companies would need to invest in community trust without clear returns; content verification providers would need to rebuild technology around AI detection rather than traditional plagiarism.

They're building advantages that can grow over time. Pump gains collective buying power as it adds customers; Acclinate builds deeper community data and trust; GPTZero improves detection accuracy with more content to analyze. These network effects and data advantages could create barriers that strengthen as companies grow.[23]

They're changing the bases of performance. Rather than competing on traditional metrics, these companies shift value propositions: Pump focuses on flexibility and cost savings; Acclinate prioritizes trust and inclusion; GPTZero emphasizes verification in a new content landscape. This redefinition of value is a pattern Christensen noted in his research.[24] They're not just moving goalposts—they're changing the game entirely.

Will they all succeed? That's far from certain—disruption is hard, and potential doesn't guarantee success. That said, our AI feels they're off to a good start, and that they're worth more than what their Pitchbook profiles might suggest. In an era where AI evaluates AI companies and the future judges itself, perhaps the most human thing we can do is recognize genuine innovation when we see it. After all, disruption has always been about seeing opportunities that others miss, and sometimes it takes a machine to point out what's hiding in plain sight.

Endnotes

[1] "Pump saves up to 60% off AWS for early stage startups, and for free (yes, that's right). Using group buying and AI, Pump automates cost savings that require no engineering effort." Y Combinator. (2024). "Pump.co: The Costco for cloud is here." https://www.ycombinator.com/companies/pump-co

[2] Christensen, C. M. (1997). The Innovator's Dilemma: When New Technologies Cause Great Firms to Fail. Harvard Business School Press.

[3] Christensen, C. M., Johnson, M. W., & Rigby, D. K. (2002). "Foundations for Growth: How to Identify and Build Disruptive New Businesses." MIT Sloan Management Review, 43(3), 22-32.

[4] "Pump uses AI & group buying to automate cloud cost optimization and savings with zero engineering input." Pump.co. (2024). https://www.pump.co/

[5] Christensen, C. M., & Bower, J. L. (1996). "Customer Power, Strategic Investment, and the Failure of Leading Firms." Strategic Management Journal, 17(3), 197-218.

[6] "We signed up over 250+ customers in the last 6 months, including over 50+ YC companies." Y Combinator. (2024). "Pump.co: The Costco for cloud is here." https://www.ycombinator.com/companies/pump-co

[7] Clinical Pharmacology & Therapeutics. (2023). "Hormonal Considerations in Clinical Trial Design and Drug Metabolism Studies." Stanford Design Lab. (2023). "Diversity in Reproductive Health Technology User Research."

[8] "For example, Albuterol, a medication often prescribed for asthma, is less effective in 67% of all Puerto Ricans and 47% of Black Americans." Acclinate. (2022). "Acclinate advances equity and diversity in clinical trials." Google Cloud Blog. https://cloud.google.com/blog/topics/startups/acclinate-advances-equity-and-diversity-in-clinical-trials

[9] "Acclinate's proprietary technology, a participation probability index... combines social determinants of health data with the individuals' engagement data to generate an index of prospective trial participants." Google for Startups.

(2024). "Acclinate - Google for Startups." https://startup.google.com/alumni/stories/acclinate/

[10] Christensen, C. M., Hall, T., Dillon, K., & Duncan, D. S. (2016). "Know Your Customers' 'Jobs to Be Done'." Harvard Business Review, 94(9), 54-62.

[11] See: "Acclinate, a Birmingham, Alabama-based healthcare technology company has secured a $7 million Series A funding round. Led by Cencora Ventures with contributions from Labcorp and Latimer Ventures." AI-TechPark. (2024). "Acclinate Raised $7M Series A to Accelerate Health Equity Revolution." https://ai-techpark.com/acclinate-raised-7m-series-a-to-accelerate-health-equity-revolution/

[12] See: Christensen, C. M., Johnson, M. W., & Sinfield, J. V. (2008). "Reinventing Your Business Model." Harvard Business Review, 86(12), 50-59.

[13] "GPTZero is the most accurate AI detector across use-cases, verified by multiple independent sources, including TechCrunch, which called us the best and most reliable AI detector after testing seven others." GPTZero. (2024). https://gptzero.me/

[14] "GPTZero grew 500% in ARR in the last six months, the founders told TechCrunch, adding that its user base has grown from 1 million to 4 million in the last 12 months." TechCrunch. (2024). "GPTZero's founders, still in their 20s, have a profitable AI detection startup, millions in the bank and a new $10M Series A." https://techcrunch.com/2024/06/13/gptzero-profitable-ai-detection-startup-10m-series-a/

[15] "While the startup may be best known for helping teachers detect AI-generated student work (in October, GPTZero landed an agreement with the American Federation of Teachers), its customer base has expanded. It now includes government procurement agencies, grant-writing organizations, hiring managers, and — especially interesting — AI training data labelers." TechCrunch. (2024). "GPTZero's founders, still in their 20s, have a profitable AI detection startup, millions in the bank and a new $10M Series A." https://techcrunch.com/2024/06/13/gptzero-profitable-ai-detection-startup-10m-series-a/

[16] Christensen, C. M., & Raynor, M. E. (2003). The Innovator's Solution: Creating and Sustaining Successful Growth. Harvard Business School Press.

[17] "Disk Defragmenter first shipped as part of Windows 95 and later shipped with Windows 98 and Windows Me, licensed from Symantec Corporation." Wikipedia. (2025). "Microsoft Drive Optimizer." https://en.wikipedia.org/wiki/Microsoft_Drive_Optimizer

[18]For another example, see: "Kevin Systrom said that Instagram's 'Stories' feature is 'definitely similar' to Snapchat's version, but argued that it did not establish that they were copying." Business Standard. (2019). "Does Instagram copy Snapchat? Insta CEO admits 'stories' similar to 'snaps'." https://www.business-standard.com/article/technology/does-instagram-copy-snapchat-insta-ceo-admits-stories-similar-to-snaps-117053100764_1.html

[19] For another example, see: "Instagram has over 200 million people using 'Stories' daily - more than even Snapchat." Business Standard. (2019). "Does Instagram copy Snapchat? Insta CEO admits 'stories' similar to 'snaps'." https://www.business-standard.com/article/technology/does-instagram-copy-snapchat-insta-ceo-admits-stories-similar-to-snaps-117053100764_1.html

[20] Christensen, C. M. (1997). The Innovator's Dilemma: When New Technologies Cause Great Firms to Fail. Harvard Business School Press.

[21] Christensen, C. M., Johnson, M. W., & Sinfield, J. V. (2008). "Reinventing Your Business Model." Harvard Business Review, 86(12), 50-59.

[22] Christensen, C. M., & Overdorf, M. (2000). "Meeting the Challenge of Disruptive Change." Harvard Business Review, 78(2), 66-77.

[23] Hagel, J., & Singer, M. (1999). "Unbundling the Corporation." Harvard Business Review, 77(2), 133-141.

[24] Christensen, C. M., Cook, S., & Hall, T. (2005). "Marketing Malpractice: The Cause and the Cure." Harvard Business Review, 83(12), 74-83.